| @@ -1,2086 +1,2086 | |||
|
|
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 | import math |
|
|
16 | 16 | |
|
|
17 | 17 | from schainpy.model.proc.jroproc_base import ProcessingUnit, MPDecorator, Operation |
|
|
18 | 18 | from schainpy.model.data.jrodata import Spectra |
|
|
19 | 19 | from schainpy.model.data.jrodata import hildebrand_sekhon |
|
|
20 | 20 | from schainpy.model.data import _noise |
|
|
21 | 21 | |
|
|
22 | 22 | from schainpy.utils import log |
|
|
23 | 23 | import matplotlib.pyplot as plt |
|
|
24 | 24 | #from scipy.optimize import curve_fit |
|
|
25 | 25 | |
|
|
26 | 26 | class SpectraProc(ProcessingUnit): |
|
|
27 | 27 | |
|
|
28 | 28 | def __init__(self): |
|
|
29 | 29 | |
|
|
30 | 30 | ProcessingUnit.__init__(self) |
|
|
31 | 31 | |
|
|
32 | 32 | self.buffer = None |
|
|
33 | 33 | self.firstdatatime = None |
|
|
34 | 34 | self.profIndex = 0 |
|
|
35 | 35 | self.dataOut = Spectra() |
|
|
36 | 36 | self.id_min = None |
|
|
37 | 37 | self.id_max = None |
|
|
38 | 38 | self.setupReq = False #Agregar a todas las unidades de proc |
|
|
39 | 39 | |
|
|
40 | 40 | def __updateSpecFromVoltage(self): |
|
|
41 | 41 | |
|
|
42 | 42 | |
|
|
43 | 43 | |
|
|
44 | 44 | self.dataOut.timeZone = self.dataIn.timeZone |
|
|
45 | 45 | self.dataOut.dstFlag = self.dataIn.dstFlag |
|
|
46 | 46 | self.dataOut.errorCount = self.dataIn.errorCount |
|
|
47 | 47 | self.dataOut.useLocalTime = self.dataIn.useLocalTime |
|
|
48 | 48 | try: |
|
|
49 | 49 | self.dataOut.processingHeaderObj = self.dataIn.processingHeaderObj.copy() |
|
|
50 | 50 | except: |
|
|
51 | 51 | pass |
|
|
52 | 52 | self.dataOut.radarControllerHeaderObj = self.dataIn.radarControllerHeaderObj.copy() |
|
|
53 | 53 | self.dataOut.systemHeaderObj = self.dataIn.systemHeaderObj.copy() |
|
|
54 | 54 | self.dataOut.channelList = self.dataIn.channelList |
|
|
55 | 55 | self.dataOut.heightList = self.dataIn.heightList |
|
|
56 | 56 | self.dataOut.dtype = numpy.dtype([('real', '<f4'), ('imag', '<f4')]) |
|
|
57 | 57 | self.dataOut.nProfiles = self.dataOut.nFFTPoints |
|
|
58 | 58 | self.dataOut.flagDiscontinuousBlock = self.dataIn.flagDiscontinuousBlock |
|
|
59 | 59 | self.dataOut.utctime = self.firstdatatime |
|
|
60 | 60 | self.dataOut.flagDecodeData = self.dataIn.flagDecodeData |
|
|
61 | 61 | self.dataOut.flagDeflipData = self.dataIn.flagDeflipData |
|
|
62 | 62 | self.dataOut.flagShiftFFT = False |
|
|
63 | 63 | self.dataOut.nCohInt = self.dataIn.nCohInt |
|
|
64 | 64 | self.dataOut.nIncohInt = 1 |
|
|
65 | 65 | self.dataOut.windowOfFilter = self.dataIn.windowOfFilter |
|
|
66 | 66 | self.dataOut.frequency = self.dataIn.frequency |
|
|
67 | 67 | self.dataOut.realtime = self.dataIn.realtime |
|
|
68 | 68 | self.dataOut.azimuth = self.dataIn.azimuth |
|
|
69 | 69 | self.dataOut.zenith = self.dataIn.zenith |
|
|
70 | 70 | self.dataOut.codeList = self.dataIn.codeList |
|
|
71 | 71 | self.dataOut.azimuthList = self.dataIn.azimuthList |
|
|
72 | 72 | self.dataOut.elevationList = self.dataIn.elevationList |
|
|
73 | 73 | |
|
|
74 | 74 | |
|
|
75 | 75 | def __getFft(self): |
|
|
76 | 76 | """ |
|
|
77 | 77 | Convierte valores de Voltaje a Spectra |
|
|
78 | 78 | |
|
|
79 | 79 | Affected: |
|
|
80 | 80 | self.dataOut.data_spc |
|
|
81 | 81 | self.dataOut.data_cspc |
|
|
82 | 82 | self.dataOut.data_dc |
|
|
83 | 83 | self.dataOut.heightList |
|
|
84 | 84 | self.profIndex |
|
|
85 | 85 | self.buffer |
|
|
86 | 86 | self.dataOut.flagNoData |
|
|
87 | 87 | """ |
|
|
88 | 88 | fft_volt = numpy.fft.fft( |
|
|
89 | 89 | self.buffer, n=self.dataOut.nFFTPoints, axis=1) |
|
|
90 | 90 | fft_volt = fft_volt.astype(numpy.dtype('complex')) |
|
|
91 | 91 | dc = fft_volt[:, 0, :] |
|
|
92 | 92 | |
|
|
93 | 93 | # calculo de self-spectra |
|
|
94 | 94 | fft_volt = numpy.fft.fftshift(fft_volt, axes=(1,)) |
|
|
95 | 95 | spc = fft_volt * numpy.conjugate(fft_volt) |
|
|
96 | 96 | spc = spc.real |
|
|
97 | 97 | |
|
|
98 | 98 | blocksize = 0 |
|
|
99 | 99 | blocksize += dc.size |
|
|
100 | 100 | blocksize += spc.size |
|
|
101 | 101 | |
|
|
102 | 102 | cspc = None |
|
|
103 | 103 | pairIndex = 0 |
|
|
104 | 104 | if self.dataOut.pairsList != None: |
|
|
105 | 105 | # calculo de cross-spectra |
|
|
106 | 106 | cspc = numpy.zeros( |
|
|
107 | 107 | (self.dataOut.nPairs, self.dataOut.nFFTPoints, self.dataOut.nHeights), dtype='complex') |
|
|
108 | 108 | for pair in self.dataOut.pairsList: |
|
|
109 | 109 | if pair[0] not in self.dataOut.channelList: |
|
|
110 | 110 | raise ValueError("Error getting CrossSpectra: pair 0 of %s is not in channelList = %s" % ( |
|
|
111 | 111 | str(pair), str(self.dataOut.channelList))) |
|
|
112 | 112 | if pair[1] not in self.dataOut.channelList: |
|
|
113 | 113 | raise ValueError("Error getting CrossSpectra: pair 1 of %s is not in channelList = %s" % ( |
|
|
114 | 114 | str(pair), str(self.dataOut.channelList))) |
|
|
115 | 115 | |
|
|
116 | 116 | cspc[pairIndex, :, :] = fft_volt[pair[0], :, :] * \ |
|
|
117 | 117 | numpy.conjugate(fft_volt[pair[1], :, :]) |
|
|
118 | 118 | pairIndex += 1 |
|
|
119 | 119 | blocksize += cspc.size |
|
|
120 | 120 | |
|
|
121 | 121 | self.dataOut.data_spc = spc |
|
|
122 | 122 | self.dataOut.data_cspc = cspc |
|
|
123 | 123 | self.dataOut.data_dc = dc |
|
|
124 | 124 | self.dataOut.blockSize = blocksize |
|
|
125 | 125 | self.dataOut.flagShiftFFT = False |
|
|
126 | 126 | |
|
|
127 | 127 | def run(self, nProfiles=None, nFFTPoints=None, pairsList=None, ippFactor=None, shift_fft=False): |
|
|
128 | 128 | #print("run spc proc") |
|
|
129 | 129 | try: |
|
|
130 | 130 | type = self.dataIn.type.decode("utf-8") |
|
|
131 | 131 | self.dataIn.type = type |
|
|
132 | 132 | except: |
|
|
133 | 133 | pass |
|
|
134 | 134 | if self.dataIn.type == "Spectra": |
|
|
135 | 135 | |
|
|
136 | 136 | try: |
|
|
137 | 137 | self.dataOut.copy(self.dataIn) |
|
|
138 | 138 | |
|
|
139 | 139 | except Exception as e: |
|
|
140 | 140 | print("Error dataIn ",e) |
|
|
141 | 141 | |
|
|
142 | 142 | if shift_fft: |
|
|
143 | 143 | #desplaza a la derecha en el eje 2 determinadas posiciones |
|
|
144 | 144 | shift = int(self.dataOut.nFFTPoints/2) |
|
|
145 | 145 | self.dataOut.data_spc = numpy.roll(self.dataOut.data_spc, shift , axis=1) |
|
|
146 | 146 | |
|
|
147 | 147 | if self.dataOut.data_cspc is not None: |
|
|
148 | 148 | #desplaza a la derecha en el eje 2 determinadas posiciones |
|
|
149 | 149 | self.dataOut.data_cspc = numpy.roll(self.dataOut.data_cspc, shift, axis=1) |
|
|
150 | 150 | if pairsList: |
|
|
151 | 151 | self.__selectPairs(pairsList) |
|
|
152 | 152 | |
|
|
153 | 153 | |
|
|
154 | 154 | elif self.dataIn.type == "Voltage": |
|
|
155 | 155 | |
|
|
156 | 156 | self.dataOut.flagNoData = True |
|
|
157 | 157 | |
|
|
158 | 158 | if nFFTPoints == None: |
|
|
159 | 159 | raise ValueError("This SpectraProc.run() need nFFTPoints input variable") |
|
|
160 | 160 | |
|
|
161 | 161 | if nProfiles == None: |
|
|
162 | 162 | nProfiles = nFFTPoints |
|
|
163 | 163 | |
|
|
164 | 164 | if ippFactor == None: |
|
|
165 | 165 | self.dataOut.ippFactor = 1 |
|
|
166 | 166 | |
|
|
167 | 167 | self.dataOut.nFFTPoints = nFFTPoints |
|
|
168 | 168 | #print(" volts ch,prof, h: ", self.dataIn.data.shape) |
|
|
169 | 169 | if self.buffer is None: |
|
|
170 | 170 | self.buffer = numpy.zeros((self.dataIn.nChannels, |
|
|
171 | 171 | nProfiles, |
|
|
172 | 172 | self.dataIn.nHeights), |
|
|
173 | 173 | dtype='complex') |
|
|
174 | 174 | |
|
|
175 | 175 | if self.dataIn.flagDataAsBlock: |
|
|
176 | 176 | nVoltProfiles = self.dataIn.data.shape[1] |
|
|
177 | 177 | |
|
|
178 | 178 | if nVoltProfiles == nProfiles: |
|
|
179 | 179 | self.buffer = self.dataIn.data.copy() |
|
|
180 | 180 | self.profIndex = nVoltProfiles |
|
|
181 | 181 | |
|
|
182 | 182 | elif nVoltProfiles < nProfiles: |
|
|
183 | 183 | |
|
|
184 | 184 | if self.profIndex == 0: |
|
|
185 | 185 | self.id_min = 0 |
|
|
186 | 186 | self.id_max = nVoltProfiles |
|
|
187 | 187 | |
|
|
188 | 188 | self.buffer[:, self.id_min:self.id_max, |
|
|
189 | 189 | :] = self.dataIn.data |
|
|
190 | 190 | self.profIndex += nVoltProfiles |
|
|
191 | 191 | self.id_min += nVoltProfiles |
|
|
192 | 192 | self.id_max += nVoltProfiles |
|
|
193 | 193 | else: |
|
|
194 | 194 | raise ValueError("The type object %s has %d profiles, it should just has %d profiles" % ( |
|
|
195 | 195 | self.dataIn.type, self.dataIn.data.shape[1], nProfiles)) |
|
|
196 | 196 | self.dataOut.flagNoData = True |
|
|
197 | 197 | else: |
|
|
198 | 198 | self.buffer[:, self.profIndex, :] = self.dataIn.data.copy() |
|
|
199 | 199 | self.profIndex += 1 |
|
|
200 | 200 | |
|
|
201 | 201 | if self.firstdatatime == None: |
|
|
202 | 202 | self.firstdatatime = self.dataIn.utctime |
|
|
203 | 203 | |
|
|
204 | 204 | if self.profIndex == nProfiles: |
|
|
205 | 205 | |
|
|
206 | 206 | self.__updateSpecFromVoltage() |
|
|
207 | 207 | |
|
|
208 | 208 | if pairsList == None: |
|
|
209 | 209 | self.dataOut.pairsList = [pair for pair in itertools.combinations(self.dataOut.channelList, 2)] |
|
|
210 | 210 | else: |
|
|
211 | 211 | self.dataOut.pairsList = pairsList |
|
|
212 | 212 | self.__getFft() |
|
|
213 | 213 | self.dataOut.flagNoData = False |
|
|
214 | 214 | self.firstdatatime = None |
|
|
215 | 215 | self.profIndex = 0 |
|
|
216 | 216 | |
|
|
217 | 217 | elif self.dataIn.type == "Parameters": |
|
|
218 | 218 | |
|
|
219 | 219 | self.dataOut.data_spc = self.dataIn.data_spc |
|
|
220 | 220 | self.dataOut.data_cspc = self.dataIn.data_cspc |
|
|
221 | 221 | self.dataOut.data_outlier = self.dataIn.data_outlier |
|
|
222 | 222 | self.dataOut.nProfiles = self.dataIn.nProfiles |
|
|
223 | 223 | self.dataOut.nIncohInt = self.dataIn.nIncohInt |
|
|
224 | 224 | self.dataOut.nFFTPoints = self.dataIn.nFFTPoints |
|
|
225 | 225 | self.dataOut.ippFactor = self.dataIn.ippFactor |
|
|
226 | 226 | self.dataOut.max_nIncohInt = self.dataIn.max_nIncohInt |
|
|
227 | 227 | self.dataOut.radarControllerHeaderObj = self.dataIn.radarControllerHeaderObj.copy() |
|
|
228 | 228 | self.dataOut.ipp = self.dataIn.ipp |
|
|
229 | 229 | #self.dataOut.abscissaList = self.dataIn.getVelRange(1) |
|
|
230 | 230 | #self.dataOut.spc_noise = self.dataIn.getNoise() |
|
|
231 | 231 | #self.dataOut.spc_range = (self.dataIn.getFreqRange(1) , self.dataIn.getAcfRange(1) , self.dataIn.getVelRange(1)) |
|
|
232 | 232 | # self.dataOut.normFactor = self.dataIn.normFactor |
|
|
233 | 233 | if hasattr(self.dataIn, 'channelList'): |
|
|
234 | 234 | self.dataOut.channelList = self.dataIn.channelList |
|
|
235 | 235 | if hasattr(self.dataIn, 'pairsList'): |
|
|
236 | 236 | self.dataOut.pairsList = self.dataIn.pairsList |
|
|
237 | 237 | self.dataOut.groupList = self.dataIn.pairsList |
|
|
238 | 238 | |
|
|
239 | 239 | self.dataOut.flagNoData = False |
|
|
240 | 240 | |
|
|
241 | 241 | if hasattr(self.dataIn, 'ChanDist'): #Distances of receiver channels |
|
|
242 | 242 | self.dataOut.ChanDist = self.dataIn.ChanDist |
|
|
243 | 243 | else: self.dataOut.ChanDist = None |
|
|
244 | 244 | |
|
|
245 | 245 | #if hasattr(self.dataIn, 'VelRange'): #Velocities range |
|
|
246 | 246 | # self.dataOut.VelRange = self.dataIn.VelRange |
|
|
247 | 247 | #else: self.dataOut.VelRange = None |
|
|
248 | 248 | |
|
|
249 | 249 | |
|
|
250 | 250 | |
|
|
251 | 251 | else: |
|
|
252 | 252 | raise ValueError("The type of input object {} is not valid".format( |
|
|
253 | 253 | self.dataIn.type)) |
|
|
254 | 254 | |
|
|
255 | 255 | |
|
|
256 | 256 | def __selectPairs(self, pairsList): |
|
|
257 | 257 | |
|
|
258 | 258 | if not pairsList: |
|
|
259 | 259 | return |
|
|
260 | 260 | |
|
|
261 | 261 | pairs = [] |
|
|
262 | 262 | pairsIndex = [] |
|
|
263 | 263 | |
|
|
264 | 264 | for pair in pairsList: |
|
|
265 | 265 | if pair[0] not in self.dataOut.channelList or pair[1] not in self.dataOut.channelList: |
|
|
266 | 266 | continue |
|
|
267 | 267 | pairs.append(pair) |
|
|
268 | 268 | pairsIndex.append(pairs.index(pair)) |
|
|
269 | 269 | |
|
|
270 | 270 | self.dataOut.data_cspc = self.dataOut.data_cspc[pairsIndex] |
|
|
271 | 271 | self.dataOut.pairsList = pairs |
|
|
272 | 272 | |
|
|
273 | 273 | return |
|
|
274 | 274 | |
|
|
275 | 275 | def selectFFTs(self, minFFT, maxFFT ): |
|
|
276 | 276 | """ |
|
|
277 | 277 | Selecciona un bloque de datos en base a un grupo de valores de puntos FFTs segun el rango |
|
|
278 | 278 | minFFT<= FFT <= maxFFT |
|
|
279 | 279 | """ |
|
|
280 | 280 | |
|
|
281 | 281 | if (minFFT > maxFFT): |
|
|
282 | 282 | raise ValueError("Error selecting heights: Height range (%d,%d) is not valid" % (minFFT, maxFFT)) |
|
|
283 | 283 | |
|
|
284 | 284 | if (minFFT < self.dataOut.getFreqRange()[0]): |
|
|
285 | 285 | minFFT = self.dataOut.getFreqRange()[0] |
|
|
286 | 286 | |
|
|
287 | 287 | if (maxFFT > self.dataOut.getFreqRange()[-1]): |
|
|
288 | 288 | maxFFT = self.dataOut.getFreqRange()[-1] |
|
|
289 | 289 | |
|
|
290 | 290 | minIndex = 0 |
|
|
291 | 291 | maxIndex = 0 |
|
|
292 | 292 | FFTs = self.dataOut.getFreqRange() |
|
|
293 | 293 | |
|
|
294 | 294 | inda = numpy.where(FFTs >= minFFT) |
|
|
295 | 295 | indb = numpy.where(FFTs <= maxFFT) |
|
|
296 | 296 | |
|
|
297 | 297 | try: |
|
|
298 | 298 | minIndex = inda[0][0] |
|
|
299 | 299 | except: |
|
|
300 | 300 | minIndex = 0 |
|
|
301 | 301 | |
|
|
302 | 302 | try: |
|
|
303 | 303 | maxIndex = indb[0][-1] |
|
|
304 | 304 | except: |
|
|
305 | 305 | maxIndex = len(FFTs) |
|
|
306 | 306 | |
|
|
307 | 307 | self.selectFFTsByIndex(minIndex, maxIndex) |
|
|
308 | 308 | |
|
|
309 | 309 | return 1 |
|
|
310 | 310 | |
|
|
311 | 311 | def getBeaconSignal(self, tauindex=0, channelindex=0, hei_ref=None): |
|
|
312 | 312 | newheis = numpy.where( |
|
|
313 | 313 | self.dataOut.heightList > self.dataOut.radarControllerHeaderObj.Taus[tauindex]) |
|
|
314 | 314 | |
|
|
315 | 315 | if hei_ref != None: |
|
|
316 | 316 | newheis = numpy.where(self.dataOut.heightList > hei_ref) |
|
|
317 | 317 | |
|
|
318 | 318 | minIndex = min(newheis[0]) |
|
|
319 | 319 | maxIndex = max(newheis[0]) |
|
|
320 | 320 | data_spc = self.dataOut.data_spc[:, :, minIndex:maxIndex + 1] |
|
|
321 | 321 | heightList = self.dataOut.heightList[minIndex:maxIndex + 1] |
|
|
322 | 322 | |
|
|
323 | 323 | # determina indices |
|
|
324 | 324 | nheis = int(self.dataOut.radarControllerHeaderObj.txB / |
|
|
325 | 325 | (self.dataOut.heightList[1] - self.dataOut.heightList[0])) |
|
|
326 | 326 | avg_dB = 10 * \ |
|
|
327 | 327 | numpy.log10(numpy.sum(data_spc[channelindex, :, :], axis=0)) |
|
|
328 | 328 | beacon_dB = numpy.sort(avg_dB)[-nheis:] |
|
|
329 | 329 | beacon_heiIndexList = [] |
|
|
330 | 330 | for val in avg_dB.tolist(): |
|
|
331 | 331 | if val >= beacon_dB[0]: |
|
|
332 | 332 | beacon_heiIndexList.append(avg_dB.tolist().index(val)) |
|
|
333 | 333 | |
|
|
334 | 334 | #data_spc = data_spc[:,:,beacon_heiIndexList] |
|
|
335 | 335 | data_cspc = None |
|
|
336 | 336 | if self.dataOut.data_cspc is not None: |
|
|
337 | 337 | data_cspc = self.dataOut.data_cspc[:, :, minIndex:maxIndex + 1] |
|
|
338 | 338 | #data_cspc = data_cspc[:,:,beacon_heiIndexList] |
|
|
339 | 339 | |
|
|
340 | 340 | data_dc = None |
|
|
341 | 341 | if self.dataOut.data_dc is not None: |
|
|
342 | 342 | data_dc = self.dataOut.data_dc[:, minIndex:maxIndex + 1] |
|
|
343 | 343 | #data_dc = data_dc[:,beacon_heiIndexList] |
|
|
344 | 344 | |
|
|
345 | 345 | self.dataOut.data_spc = data_spc |
|
|
346 | 346 | self.dataOut.data_cspc = data_cspc |
|
|
347 | 347 | self.dataOut.data_dc = data_dc |
|
|
348 | 348 | self.dataOut.heightList = heightList |
|
|
349 | 349 | self.dataOut.beacon_heiIndexList = beacon_heiIndexList |
|
|
350 | 350 | |
|
|
351 | 351 | return 1 |
|
|
352 | 352 | |
|
|
353 | 353 | def selectFFTsByIndex(self, minIndex, maxIndex): |
|
|
354 | 354 | """ |
|
|
355 | 355 | |
|
|
356 | 356 | """ |
|
|
357 | 357 | |
|
|
358 | 358 | if (minIndex < 0) or (minIndex > maxIndex): |
|
|
359 | 359 | raise ValueError("Error selecting heights: Index range (%d,%d) is not valid" % (minIndex, maxIndex)) |
|
|
360 | 360 | |
|
|
361 | 361 | if (maxIndex >= self.dataOut.nProfiles): |
|
|
362 | 362 | maxIndex = self.dataOut.nProfiles-1 |
|
|
363 | 363 | |
|
|
364 | 364 | #Spectra |
|
|
365 | 365 | data_spc = self.dataOut.data_spc[:,minIndex:maxIndex+1,:] |
|
|
366 | 366 | |
|
|
367 | 367 | data_cspc = None |
|
|
368 | 368 | if self.dataOut.data_cspc is not None: |
|
|
369 | 369 | data_cspc = self.dataOut.data_cspc[:,minIndex:maxIndex+1,:] |
|
|
370 | 370 | |
|
|
371 | 371 | data_dc = None |
|
|
372 | 372 | if self.dataOut.data_dc is not None: |
|
|
373 | 373 | data_dc = self.dataOut.data_dc[minIndex:maxIndex+1,:] |
|
|
374 | 374 | |
|
|
375 | 375 | self.dataOut.data_spc = data_spc |
|
|
376 | 376 | self.dataOut.data_cspc = data_cspc |
|
|
377 | 377 | self.dataOut.data_dc = data_dc |
|
|
378 | 378 | |
|
|
379 | 379 | self.dataOut.ippSeconds = self.dataOut.ippSeconds*(self.dataOut.nFFTPoints / numpy.shape(data_cspc)[1]) |
|
|
380 | 380 | self.dataOut.nFFTPoints = numpy.shape(data_cspc)[1] |
|
|
381 | 381 | self.dataOut.profilesPerBlock = numpy.shape(data_cspc)[1] |
|
|
382 | 382 | |
|
|
383 | 383 | return 1 |
|
|
384 | 384 | |
|
|
385 | 385 | def getNoise(self, minHei=None, maxHei=None, minVel=None, maxVel=None): |
|
|
386 | 386 | # validacion de rango |
|
|
387 | 387 | if minHei == None: |
|
|
388 | 388 | minHei = self.dataOut.heightList[0] |
|
|
389 | 389 | |
|
|
390 | 390 | if maxHei == None: |
|
|
391 | 391 | maxHei = self.dataOut.heightList[-1] |
|
|
392 | 392 | |
|
|
393 | 393 | if (minHei < self.dataOut.heightList[0]) or (minHei > maxHei): |
|
|
394 | 394 | print('minHei: %.2f is out of the heights range' % (minHei)) |
|
|
395 | 395 | print('minHei is setting to %.2f' % (self.dataOut.heightList[0])) |
|
|
396 | 396 | minHei = self.dataOut.heightList[0] |
|
|
397 | 397 | |
|
|
398 | 398 | if (maxHei > self.dataOut.heightList[-1]) or (maxHei < minHei): |
|
|
399 | 399 | print('maxHei: %.2f is out of the heights range' % (maxHei)) |
|
|
400 | 400 | print('maxHei is setting to %.2f' % (self.dataOut.heightList[-1])) |
|
|
401 | 401 | maxHei = self.dataOut.heightList[-1] |
|
|
402 | 402 | |
|
|
403 | 403 | # validacion de velocidades |
|
|
404 | 404 | velrange = self.dataOut.getVelRange(1) |
|
|
405 | 405 | |
|
|
406 | 406 | if minVel == None: |
|
|
407 | 407 | minVel = velrange[0] |
|
|
408 | 408 | |
|
|
409 | 409 | if maxVel == None: |
|
|
410 | 410 | maxVel = velrange[-1] |
|
|
411 | 411 | |
|
|
412 | 412 | if (minVel < velrange[0]) or (minVel > maxVel): |
|
|
413 | 413 | print('minVel: %.2f is out of the velocity range' % (minVel)) |
|
|
414 | 414 | print('minVel is setting to %.2f' % (velrange[0])) |
|
|
415 | 415 | minVel = velrange[0] |
|
|
416 | 416 | |
|
|
417 | 417 | if (maxVel > velrange[-1]) or (maxVel < minVel): |
|
|
418 | 418 | print('maxVel: %.2f is out of the velocity range' % (maxVel)) |
|
|
419 | 419 | print('maxVel is setting to %.2f' % (velrange[-1])) |
|
|
420 | 420 | maxVel = velrange[-1] |
|
|
421 | 421 | |
|
|
422 | 422 | # seleccion de indices para rango |
|
|
423 | 423 | minIndex = 0 |
|
|
424 | 424 | maxIndex = 0 |
|
|
425 | 425 | heights = self.dataOut.heightList |
|
|
426 | 426 | |
|
|
427 | 427 | inda = numpy.where(heights >= minHei) |
|
|
428 | 428 | indb = numpy.where(heights <= maxHei) |
|
|
429 | 429 | |
|
|
430 | 430 | try: |
|
|
431 | 431 | minIndex = inda[0][0] |
|
|
432 | 432 | except: |
|
|
433 | 433 | minIndex = 0 |
|
|
434 | 434 | |
|
|
435 | 435 | try: |
|
|
436 | 436 | maxIndex = indb[0][-1] |
|
|
437 | 437 | except: |
|
|
438 | 438 | maxIndex = len(heights) |
|
|
439 | 439 | |
|
|
440 | 440 | if (minIndex < 0) or (minIndex > maxIndex): |
|
|
441 | 441 | raise ValueError("some value in (%d,%d) is not valid" % ( |
|
|
442 | 442 | minIndex, maxIndex)) |
|
|
443 | 443 | |
|
|
444 | 444 | if (maxIndex >= self.dataOut.nHeights): |
|
|
445 | 445 | maxIndex = self.dataOut.nHeights - 1 |
|
|
446 | 446 | |
|
|
447 | 447 | # seleccion de indices para velocidades |
|
|
448 | 448 | indminvel = numpy.where(velrange >= minVel) |
|
|
449 | 449 | indmaxvel = numpy.where(velrange <= maxVel) |
|
|
450 | 450 | try: |
|
|
451 | 451 | minIndexVel = indminvel[0][0] |
|
|
452 | 452 | except: |
|
|
453 | 453 | minIndexVel = 0 |
|
|
454 | 454 | |
|
|
455 | 455 | try: |
|
|
456 | 456 | maxIndexVel = indmaxvel[0][-1] |
|
|
457 | 457 | except: |
|
|
458 | 458 | maxIndexVel = len(velrange) |
|
|
459 | 459 | |
|
|
460 | 460 | # seleccion del espectro |
|
|
461 | 461 | data_spc = self.dataOut.data_spc[:, |
|
|
462 | 462 | minIndexVel:maxIndexVel + 1, minIndex:maxIndex + 1] |
|
|
463 | 463 | # estimacion de ruido |
|
|
464 | 464 | noise = numpy.zeros(self.dataOut.nChannels) |
|
|
465 | 465 | |
|
|
466 | 466 | for channel in range(self.dataOut.nChannels): |
|
|
467 | 467 | daux = data_spc[channel, :, :] |
|
|
468 | 468 | sortdata = numpy.sort(daux, axis=None) |
|
|
469 | 469 | noise[channel] = hildebrand_sekhon(sortdata, self.dataOut.nIncohInt) |
|
|
470 | 470 | |
|
|
471 | 471 | self.dataOut.noise_estimation = noise.copy() |
|
|
472 | 472 | |
|
|
473 | 473 | return 1 |
|
|
474 | 474 | |
|
|
475 | 475 | class removeDC(Operation): |
|
|
476 | 476 | |
|
|
477 | 477 | def run(self, dataOut, mode=2): |
|
|
478 | 478 | self.dataOut = dataOut |
|
|
479 | 479 | jspectra = self.dataOut.data_spc |
|
|
480 | 480 | jcspectra = self.dataOut.data_cspc |
|
|
481 | 481 | |
|
|
482 | 482 | num_chan = jspectra.shape[0] |
|
|
483 | 483 | num_hei = jspectra.shape[2] |
|
|
484 | 484 | |
|
|
485 | 485 | if jcspectra is not None: |
|
|
486 | 486 | jcspectraExist = True |
|
|
487 | 487 | num_pairs = jcspectra.shape[0] |
|
|
488 | 488 | else: |
|
|
489 | 489 | jcspectraExist = False |
|
|
490 | 490 | |
|
|
491 | 491 | freq_dc = int(jspectra.shape[1] / 2) |
|
|
492 | 492 | ind_vel = numpy.array([-2, -1, 1, 2]) + freq_dc |
|
|
493 | 493 | ind_vel = ind_vel.astype(int) |
|
|
494 | 494 | |
|
|
495 | 495 | if ind_vel[0] < 0: |
|
|
496 | 496 | ind_vel[list(range(0, 1))] = ind_vel[list(range(0, 1))] + self.num_prof |
|
|
497 | 497 | |
|
|
498 | 498 | if mode == 1: |
|
|
499 | 499 | jspectra[:, freq_dc, :] = ( |
|
|
500 | 500 | jspectra[:, ind_vel[1], :] + jspectra[:, ind_vel[2], :]) / 2 # CORRECCION |
|
|
501 | 501 | |
|
|
502 | 502 | if jcspectraExist: |
|
|
503 | 503 | jcspectra[:, freq_dc, :] = ( |
|
|
504 | 504 | jcspectra[:, ind_vel[1], :] + jcspectra[:, ind_vel[2], :]) / 2 |
|
|
505 | 505 | |
|
|
506 | 506 | if mode == 2: |
|
|
507 | 507 | |
|
|
508 | 508 | vel = numpy.array([-2, -1, 1, 2]) |
|
|
509 | 509 | xx = numpy.zeros([4, 4]) |
|
|
510 | 510 | |
|
|
511 | 511 | for fil in range(4): |
|
|
512 | 512 | xx[fil, :] = vel[fil]**numpy.asarray(list(range(4))) |
|
|
513 | 513 | |
|
|
514 | 514 | xx_inv = numpy.linalg.inv(xx) |
|
|
515 | 515 | xx_aux = xx_inv[0, :] |
|
|
516 | 516 | |
|
|
517 | 517 | for ich in range(num_chan): |
|
|
518 | 518 | yy = jspectra[ich, ind_vel, :] |
|
|
519 | 519 | jspectra[ich, freq_dc, :] = numpy.dot(xx_aux, yy) |
|
|
520 | 520 | |
|
|
521 | 521 | junkid = jspectra[ich, freq_dc, :] <= 0 |
|
|
522 | 522 | cjunkid = sum(junkid) |
|
|
523 | 523 | |
|
|
524 | 524 | if cjunkid.any(): |
|
|
525 | 525 | jspectra[ich, freq_dc, junkid.nonzero()] = ( |
|
|
526 | 526 | jspectra[ich, ind_vel[1], junkid] + jspectra[ich, ind_vel[2], junkid]) / 2 |
|
|
527 | 527 | |
|
|
528 | 528 | if jcspectraExist: |
|
|
529 | 529 | for ip in range(num_pairs): |
|
|
530 | 530 | yy = jcspectra[ip, ind_vel, :] |
|
|
531 | 531 | jcspectra[ip, freq_dc, :] = numpy.dot(xx_aux, yy) |
|
|
532 | 532 | |
|
|
533 | 533 | self.dataOut.data_spc = jspectra |
|
|
534 | 534 | self.dataOut.data_cspc = jcspectra |
|
|
535 | 535 | |
|
|
536 | 536 | return self.dataOut |
|
|
537 | 537 | |
|
|
538 | 538 | class getNoiseB(Operation): |
|
|
539 | 539 | |
|
|
540 | 540 | __slots__ =('offset','warnings', 'isConfig', 'minIndex','maxIndex','minIndexFFT','maxIndexFFT') |
|
|
541 | 541 | def __init__(self): |
|
|
542 | 542 | |
|
|
543 | 543 | Operation.__init__(self) |
|
|
544 | 544 | self.isConfig = False |
|
|
545 | 545 | |
|
|
546 | 546 | def setup(self, offset=None, minHei=None, maxHei=None,minVel=None, maxVel=None, minFreq= None, maxFreq=None, warnings=False): |
|
|
547 | 547 | |
|
|
548 | 548 | self.warnings = warnings |
|
|
549 | 549 | if minHei == None: |
|
|
550 | 550 | minHei = self.dataOut.heightList[0] |
|
|
551 | 551 | |
|
|
552 | 552 | if maxHei == None: |
|
|
553 | 553 | maxHei = self.dataOut.heightList[-1] |
|
|
554 | 554 | |
|
|
555 | 555 | if (minHei < self.dataOut.heightList[0]) or (minHei > maxHei): |
|
|
556 | 556 | if self.warnings: |
|
|
557 | 557 | print('minHei: %.2f is out of the heights range' % (minHei)) |
|
|
558 | 558 | print('minHei is setting to %.2f' % (self.dataOut.heightList[0])) |
|
|
559 | 559 | minHei = self.dataOut.heightList[0] |
|
|
560 | 560 | |
|
|
561 | 561 | if (maxHei > self.dataOut.heightList[-1]) or (maxHei < minHei): |
|
|
562 | 562 | if self.warnings: |
|
|
563 | 563 | print('maxHei: %.2f is out of the heights range' % (maxHei)) |
|
|
564 | 564 | print('maxHei is setting to %.2f' % (self.dataOut.heightList[-1])) |
|
|
565 | 565 | maxHei = self.dataOut.heightList[-1] |
|
|
566 | 566 | |
|
|
567 | 567 | |
|
|
568 | 568 | #indices relativos a los puntos de fft, puede ser de acuerdo a velocidad o frecuencia |
|
|
569 | 569 | minIndexFFT = 0 |
|
|
570 | 570 | maxIndexFFT = 0 |
|
|
571 | 571 | # validacion de velocidades |
|
|
572 | 572 | indminPoint = None |
|
|
573 | 573 | indmaxPoint = None |
|
|
574 | 574 | if self.dataOut.type == 'Spectra': |
|
|
575 | 575 | if minVel == None and maxVel == None : |
|
|
576 | 576 | |
|
|
577 | 577 | freqrange = self.dataOut.getFreqRange(1) |
|
|
578 | 578 | |
|
|
579 | 579 | if minFreq == None: |
|
|
580 | 580 | minFreq = freqrange[0] |
|
|
581 | 581 | |
|
|
582 | 582 | if maxFreq == None: |
|
|
583 | 583 | maxFreq = freqrange[-1] |
|
|
584 | 584 | |
|
|
585 | 585 | if (minFreq < freqrange[0]) or (minFreq > maxFreq): |
|
|
586 | 586 | if self.warnings: |
|
|
587 | 587 | print('minFreq: %.2f is out of the frequency range' % (minFreq)) |
|
|
588 | 588 | print('minFreq is setting to %.2f' % (freqrange[0])) |
|
|
589 | 589 | minFreq = freqrange[0] |
|
|
590 | 590 | |
|
|
591 | 591 | if (maxFreq > freqrange[-1]) or (maxFreq < minFreq): |
|
|
592 | 592 | if self.warnings: |
|
|
593 | 593 | print('maxFreq: %.2f is out of the frequency range' % (maxFreq)) |
|
|
594 | 594 | print('maxFreq is setting to %.2f' % (freqrange[-1])) |
|
|
595 | 595 | maxFreq = freqrange[-1] |
|
|
596 | 596 | |
|
|
597 | 597 | indminPoint = numpy.where(freqrange >= minFreq) |
|
|
598 | 598 | indmaxPoint = numpy.where(freqrange <= maxFreq) |
|
|
599 | 599 | |
|
|
600 | 600 | else: |
|
|
601 | 601 | |
|
|
602 | 602 | velrange = self.dataOut.getVelRange(1) |
|
|
603 | 603 | |
|
|
604 | 604 | if minVel == None: |
|
|
605 | 605 | minVel = velrange[0] |
|
|
606 | 606 | |
|
|
607 | 607 | if maxVel == None: |
|
|
608 | 608 | maxVel = velrange[-1] |
|
|
609 | 609 | |
|
|
610 | 610 | if (minVel < velrange[0]) or (minVel > maxVel): |
|
|
611 | 611 | if self.warnings: |
|
|
612 | 612 | print('minVel: %.2f is out of the velocity range' % (minVel)) |
|
|
613 | 613 | print('minVel is setting to %.2f' % (velrange[0])) |
|
|
614 | 614 | minVel = velrange[0] |
|
|
615 | 615 | |
|
|
616 | 616 | if (maxVel > velrange[-1]) or (maxVel < minVel): |
|
|
617 | 617 | if self.warnings: |
|
|
618 | 618 | print('maxVel: %.2f is out of the velocity range' % (maxVel)) |
|
|
619 | 619 | print('maxVel is setting to %.2f' % (velrange[-1])) |
|
|
620 | 620 | maxVel = velrange[-1] |
|
|
621 | 621 | |
|
|
622 | 622 | indminPoint = numpy.where(velrange >= minVel) |
|
|
623 | 623 | indmaxPoint = numpy.where(velrange <= maxVel) |
|
|
624 | 624 | |
|
|
625 | 625 | |
|
|
626 | 626 | # seleccion de indices para rango |
|
|
627 | 627 | minIndex = 0 |
|
|
628 | 628 | maxIndex = 0 |
|
|
629 | 629 | heights = self.dataOut.heightList |
|
|
630 | 630 | |
|
|
631 | 631 | inda = numpy.where(heights >= minHei) |
|
|
632 | 632 | indb = numpy.where(heights <= maxHei) |
|
|
633 | 633 | |
|
|
634 | 634 | try: |
|
|
635 | 635 | minIndex = inda[0][0] |
|
|
636 | 636 | except: |
|
|
637 | 637 | minIndex = 0 |
|
|
638 | 638 | |
|
|
639 | 639 | try: |
|
|
640 | 640 | maxIndex = indb[0][-1] |
|
|
641 | 641 | except: |
|
|
642 | 642 | maxIndex = len(heights) |
|
|
643 | 643 | |
|
|
644 | 644 | if (minIndex < 0) or (minIndex > maxIndex): |
|
|
645 | 645 | raise ValueError("some value in (%d,%d) is not valid" % ( |
|
|
646 | 646 | minIndex, maxIndex)) |
|
|
647 | 647 | |
|
|
648 | 648 | if (maxIndex >= self.dataOut.nHeights): |
|
|
649 | 649 | maxIndex = self.dataOut.nHeights - 1 |
|
|
650 | 650 | #############################################################3 |
|
|
651 | 651 | # seleccion de indices para velocidades |
|
|
652 | 652 | if self.dataOut.type == 'Spectra': |
|
|
653 | 653 | try: |
|
|
654 | 654 | minIndexFFT = indminPoint[0][0] |
|
|
655 | 655 | except: |
|
|
656 | 656 | minIndexFFT = 0 |
|
|
657 | 657 | |
|
|
658 | 658 | try: |
|
|
659 | 659 | maxIndexFFT = indmaxPoint[0][-1] |
|
|
660 | 660 | except: |
|
|
661 | 661 | maxIndexFFT = len( self.dataOut.getFreqRange(1)) |
|
|
662 | 662 | |
|
|
663 | 663 | self.minIndex, self.maxIndex, self.minIndexFFT, self.maxIndexFFT = minIndex, maxIndex, minIndexFFT, maxIndexFFT |
|
|
664 | 664 | self.isConfig = True |
|
|
665 | 665 | if offset!=None: |
|
|
666 | 666 | self.offset = 10**(offset/10) |
|
|
667 | 667 | #print("config getNoiseB Done") |
|
|
668 | 668 | |
|
|
669 | 669 | def run(self, dataOut, offset=None, minHei=None, maxHei=None,minVel=None, maxVel=None, minFreq= None, maxFreq=None, warnings=False): |
|
|
670 | 670 | self.dataOut = dataOut |
|
|
671 | 671 | |
|
|
672 | 672 | if not self.isConfig: |
|
|
673 | 673 | self.setup(offset, minHei, maxHei,minVel, maxVel, minFreq, maxFreq, warnings) |
|
|
674 | 674 | |
|
|
675 | 675 | self.dataOut.noise_estimation = None |
|
|
676 | 676 | noise = None |
|
|
677 | 677 | if self.dataOut.type == 'Voltage': |
|
|
678 | 678 | noise = self.dataOut.getNoise(ymin_index=self.minIndex, ymax_index=self.maxIndex) |
|
|
679 | 679 | #print(minIndex, maxIndex,minIndexVel, maxIndexVel) |
|
|
680 | 680 | elif self.dataOut.type == 'Spectra': |
|
|
681 | 681 | #print(self.minIndex, self.maxIndex,self.minIndexFFT, self.maxIndexFFT, self.dataOut.nIncohInt) |
|
|
682 | 682 | noise = numpy.zeros( self.dataOut.nChannels) |
|
|
683 | 683 | norm = 1 |
|
|
684 | 684 | for channel in range( self.dataOut.nChannels): |
|
|
685 | 685 | if not hasattr(self.dataOut.nIncohInt,'__len__'): |
|
|
686 | 686 | norm = 1 |
|
|
687 | 687 | else: |
|
|
688 | 688 | norm = self.dataOut.max_nIncohInt/self.dataOut.nIncohInt[channel, self.minIndex:self.maxIndex] |
|
|
689 | 689 | #print("norm nIncoh: ", norm ,self.dataOut.data_spc.shape) |
|
|
690 | 690 | daux = self.dataOut.data_spc[channel,self.minIndexFFT:self.maxIndexFFT, self.minIndex:self.maxIndex] |
|
|
691 | 691 | daux = numpy.multiply(daux, norm) |
|
|
692 | 692 | #print("offset: ", self.offset, 10*numpy.log10(self.offset)) |
|
|
693 | 693 | #noise[channel] = self.getNoiseByMean(daux)/self.offset |
|
|
694 | 694 | #print(daux.shape, daux) |
|
|
695 | 695 | noise[channel] = self.getNoiseByHS(daux, self.dataOut.max_nIncohInt)/self.offset |
|
|
696 | 696 | |
|
|
697 | 697 | # data = numpy.mean(daux,axis=1) |
|
|
698 | 698 | # sortdata = numpy.sort(data, axis=None) |
|
|
699 | 699 | # noise[channel] = _noise.hildebrand_sekhon(sortdata, self.dataOut.max_nIncohInt)/self.offset |
|
|
700 | 700 | |
|
|
701 | 701 | #noise = self.dataOut.getNoise(xmin_index=self.minIndexFFT, xmax_index=self.maxIndexFFT, ymin_index=self.minIndex, ymax_index=self.maxIndex) |
|
|
702 | 702 | else: |
|
|
703 | 703 | noise = self.dataOut.getNoise(xmin_index=self.minIndexFFT, xmax_index=self.maxIndexFFT, ymin_index=self.minIndex, ymax_index=self.maxIndex) |
|
|
704 | 704 | self.dataOut.noise_estimation = noise.copy() # dataOut.noise |
|
|
705 | 705 | #print("2: ",10*numpy.log10(self.dataOut.noise_estimation/64)) |
|
|
706 | 706 | |
|
|
707 | 707 | #print(self.dataOut.flagNoData) |
|
|
708 | print("getNoise Done") | |
|
|
708 | print("getNoise Done", noise) | |
|
|
709 | 709 | return self.dataOut |
|
|
710 | 710 | |
|
|
711 | 711 | def getNoiseByMean(self,data): |
|
|
712 | 712 | #data debe estar ordenado |
|
|
713 | 713 | data = numpy.mean(data,axis=1) |
|
|
714 | 714 | sortdata = numpy.sort(data, axis=None) |
|
|
715 | 715 | #sortID=data.argsort() |
|
|
716 | 716 | #print(data.shape) |
|
|
717 | 717 | |
|
|
718 | 718 | pnoise = None |
|
|
719 | 719 | j = 0 |
|
|
720 | 720 | |
|
|
721 | 721 | mean = numpy.mean(sortdata) |
|
|
722 | 722 | min = numpy.min(sortdata) |
|
|
723 | 723 | delta = mean - min |
|
|
724 | 724 | indexes = numpy.where(sortdata > (mean+delta))[0] #only array of indexes |
|
|
725 | 725 | #print(len(indexes)) |
|
|
726 | 726 | if len(indexes)==0: |
|
|
727 | 727 | pnoise = numpy.mean(sortdata) |
|
|
728 | 728 | else: |
|
|
729 | 729 | j = indexes[0] |
|
|
730 | 730 | pnoise = numpy.mean(sortdata[0:j]) |
|
|
731 | 731 | |
|
|
732 | 732 | # from matplotlib import pyplot as plt |
|
|
733 | 733 | # plt.plot(sortdata) |
|
|
734 | 734 | # plt.vlines(j,(pnoise-delta),(pnoise+delta), color='r') |
|
|
735 | 735 | # plt.show() |
|
|
736 | 736 | #print("noise: ", 10*numpy.log10(pnoise)) |
|
|
737 | 737 | return pnoise |
|
|
738 | 738 | |
|
|
739 | 739 | def getNoiseByHS(self,data, navg): |
|
|
740 | 740 | #data debe estar ordenado |
|
|
741 | 741 | #data = numpy.mean(data,axis=1) |
|
|
742 | 742 | sortdata = numpy.sort(data, axis=None) |
|
|
743 | 743 | |
|
|
744 | 744 | lenOfData = len(sortdata) |
|
|
745 | 745 | nums_min = lenOfData*0.05 |
|
|
746 | 746 | |
|
|
747 | 747 | if nums_min <= 5: |
|
|
748 | 748 | |
|
|
749 | 749 | nums_min = 5 |
|
|
750 | 750 | |
|
|
751 | 751 | sump = 0. |
|
|
752 | 752 | sumq = 0. |
|
|
753 | 753 | |
|
|
754 | 754 | j = 0 |
|
|
755 | 755 | cont = 1 |
|
|
756 | 756 | |
|
|
757 | 757 | while((cont == 1)and(j < lenOfData)): |
|
|
758 | 758 | |
|
|
759 | 759 | sump += sortdata[j] |
|
|
760 | 760 | sumq += sortdata[j]**2 |
|
|
761 | 761 | #sumq -= sump**2 |
|
|
762 | 762 | if j > nums_min: |
|
|
763 | 763 | rtest = float(j)/(j-1) + 1.0/0.1 |
|
|
764 | 764 | #if ((sumq*j) > (sump**2)): |
|
|
765 | 765 | if ((sumq*j) > (rtest*sump**2)): |
|
|
766 | 766 | j = j - 1 |
|
|
767 | 767 | sump = sump - sortdata[j] |
|
|
768 | 768 | sumq = sumq - sortdata[j]**2 |
|
|
769 | 769 | cont = 0 |
|
|
770 | 770 | |
|
|
771 | 771 | j += 1 |
|
|
772 | 772 | |
|
|
773 | 773 | lnoise = sump / j |
|
|
774 | 774 | |
|
|
775 | 775 | return lnoise |
|
|
776 | 776 | |
|
|
777 | 777 | |
|
|
778 | 778 | |
|
|
779 | 779 | def fit_func( x, a0, a1, a2): #, a3, a4, a5): |
|
|
780 | 780 | z = (x - a1) / a2 |
|
|
781 | 781 | y = a0 * numpy.exp(-z**2 / a2) #+ a3 + a4 * x + a5 * x**2 |
|
|
782 | 782 | return y |
|
|
783 | 783 | |
|
|
784 | 784 | |
|
|
785 | 785 | class CleanRayleigh(Operation): |
|
|
786 | 786 | |
|
|
787 | 787 | def __init__(self): |
|
|
788 | 788 | |
|
|
789 | 789 | Operation.__init__(self) |
|
|
790 | 790 | self.i=0 |
|
|
791 | 791 | self.isConfig = False |
|
|
792 | 792 | self.__dataReady = False |
|
|
793 | 793 | self.__profIndex = 0 |
|
|
794 | 794 | self.byTime = False |
|
|
795 | 795 | self.byProfiles = False |
|
|
796 | 796 | |
|
|
797 | 797 | self.bloques = None |
|
|
798 | 798 | self.bloque0 = None |
|
|
799 | 799 | |
|
|
800 | 800 | self.index = 0 |
|
|
801 | 801 | |
|
|
802 | 802 | self.buffer = 0 |
|
|
803 | 803 | self.buffer2 = 0 |
|
|
804 | 804 | self.buffer3 = 0 |
|
|
805 | 805 | |
|
|
806 | 806 | |
|
|
807 | 807 | def setup(self,dataOut,min_hei,max_hei,n, timeInterval,factor_stdv): |
|
|
808 | 808 | |
|
|
809 | 809 | self.nChannels = dataOut.nChannels |
|
|
810 | 810 | self.nProf = dataOut.nProfiles |
|
|
811 | 811 | self.nPairs = dataOut.data_cspc.shape[0] |
|
|
812 | 812 | self.pairsArray = numpy.array(dataOut.pairsList) |
|
|
813 | 813 | self.spectra = dataOut.data_spc |
|
|
814 | 814 | self.cspectra = dataOut.data_cspc |
|
|
815 | 815 | self.heights = dataOut.heightList #alturas totales |
|
|
816 | 816 | self.nHeights = len(self.heights) |
|
|
817 | 817 | self.min_hei = min_hei |
|
|
818 | 818 | self.max_hei = max_hei |
|
|
819 | 819 | if (self.min_hei == None): |
|
|
820 | 820 | self.min_hei = 0 |
|
|
821 | 821 | if (self.max_hei == None): |
|
|
822 | 822 | self.max_hei = dataOut.heightList[-1] |
|
|
823 | 823 | self.hval = ((self.max_hei>=self.heights) & (self.heights >= self.min_hei)).nonzero() |
|
|
824 | 824 | self.heightsClean = self.heights[self.hval] #alturas filtradas |
|
|
825 | 825 | self.hval = self.hval[0] # forma (N,), an solo N elementos -> Indices de alturas |
|
|
826 | 826 | self.nHeightsClean = len(self.heightsClean) |
|
|
827 | 827 | self.channels = dataOut.channelList |
|
|
828 | 828 | self.nChan = len(self.channels) |
|
|
829 | 829 | self.nIncohInt = dataOut.nIncohInt |
|
|
830 | 830 | self.__initime = dataOut.utctime |
|
|
831 | 831 | self.maxAltInd = self.hval[-1]+1 |
|
|
832 | 832 | self.minAltInd = self.hval[0] |
|
|
833 | 833 | |
|
|
834 | 834 | self.crosspairs = dataOut.pairsList |
|
|
835 | 835 | self.nPairs = len(self.crosspairs) |
|
|
836 | 836 | self.normFactor = dataOut.normFactor |
|
|
837 | 837 | self.nFFTPoints = dataOut.nFFTPoints |
|
|
838 | 838 | self.ippSeconds = dataOut.ippSeconds |
|
|
839 | 839 | self.currentTime = self.__initime |
|
|
840 | 840 | self.pairsArray = numpy.array(dataOut.pairsList) |
|
|
841 | 841 | self.factor_stdv = factor_stdv |
|
|
842 | 842 | |
|
|
843 | 843 | if n != None : |
|
|
844 | 844 | self.byProfiles = True |
|
|
845 | 845 | self.nIntProfiles = n |
|
|
846 | 846 | else: |
|
|
847 | 847 | self.__integrationtime = timeInterval |
|
|
848 | 848 | |
|
|
849 | 849 | self.__dataReady = False |
|
|
850 | 850 | self.isConfig = True |
|
|
851 | 851 | |
|
|
852 | 852 | |
|
|
853 | 853 | |
|
|
854 | 854 | def run(self, dataOut,min_hei=None,max_hei=None, n=None, timeInterval=10,factor_stdv=2.5): |
|
|
855 | 855 | #print("runing cleanRayleigh") |
|
|
856 | 856 | if not self.isConfig : |
|
|
857 | 857 | |
|
|
858 | 858 | self.setup(dataOut, min_hei,max_hei,n,timeInterval,factor_stdv) |
|
|
859 | 859 | |
|
|
860 | 860 | tini=dataOut.utctime |
|
|
861 | 861 | |
|
|
862 | 862 | if self.byProfiles: |
|
|
863 | 863 | if self.__profIndex == self.nIntProfiles: |
|
|
864 | 864 | self.__dataReady = True |
|
|
865 | 865 | else: |
|
|
866 | 866 | if (tini - self.__initime) >= self.__integrationtime: |
|
|
867 | 867 | |
|
|
868 | 868 | self.__dataReady = True |
|
|
869 | 869 | self.__initime = tini |
|
|
870 | 870 | |
|
|
871 | 871 | #if (tini.tm_min % 2) == 0 and (tini.tm_sec < 5 and self.fint==0): |
|
|
872 | 872 | |
|
|
873 | 873 | if self.__dataReady: |
|
|
874 | 874 | |
|
|
875 | 875 | self.__profIndex = 0 |
|
|
876 | 876 | jspc = self.buffer |
|
|
877 | 877 | jcspc = self.buffer2 |
|
|
878 | 878 | #jnoise = self.buffer3 |
|
|
879 | 879 | self.buffer = dataOut.data_spc |
|
|
880 | 880 | self.buffer2 = dataOut.data_cspc |
|
|
881 | 881 | #self.buffer3 = dataOut.noise |
|
|
882 | 882 | self.currentTime = dataOut.utctime |
|
|
883 | 883 | if numpy.any(jspc) : |
|
|
884 | 884 | #print( jspc.shape, jcspc.shape) |
|
|
885 | 885 | jspc = numpy.reshape(jspc,(int(len(jspc)/self.nChannels),self.nChannels,self.nFFTPoints,self.nHeights)) |
|
|
886 | 886 | try: |
|
|
887 | 887 | jcspc= numpy.reshape(jcspc,(int(len(jcspc)/self.nPairs),self.nPairs,self.nFFTPoints,self.nHeights)) |
|
|
888 | 888 | except: |
|
|
889 | 889 | print("no cspc") |
|
|
890 | 890 | self.__dataReady = False |
|
|
891 | 891 | #print( jspc.shape, jcspc.shape) |
|
|
892 | 892 | dataOut.flagNoData = False |
|
|
893 | 893 | else: |
|
|
894 | 894 | dataOut.flagNoData = True |
|
|
895 | 895 | self.__dataReady = False |
|
|
896 | 896 | return dataOut |
|
|
897 | 897 | else: |
|
|
898 | 898 | #print( len(self.buffer)) |
|
|
899 | 899 | if numpy.any(self.buffer): |
|
|
900 | 900 | self.buffer = numpy.concatenate((self.buffer,dataOut.data_spc), axis=0) |
|
|
901 | 901 | try: |
|
|
902 | 902 | self.buffer2 = numpy.concatenate((self.buffer2,dataOut.data_cspc), axis=0) |
|
|
903 | 903 | self.buffer3 += dataOut.data_dc |
|
|
904 | 904 | except: |
|
|
905 | 905 | pass |
|
|
906 | 906 | else: |
|
|
907 | 907 | self.buffer = dataOut.data_spc |
|
|
908 | 908 | self.buffer2 = dataOut.data_cspc |
|
|
909 | 909 | self.buffer3 = dataOut.data_dc |
|
|
910 | 910 | #print self.index, self.fint |
|
|
911 | 911 | #print self.buffer2.shape |
|
|
912 | 912 | dataOut.flagNoData = True ## NOTE: ?? revisar LUEGO |
|
|
913 | 913 | self.__profIndex += 1 |
|
|
914 | 914 | return dataOut ## NOTE: REV |
|
|
915 | 915 | |
|
|
916 | 916 | |
|
|
917 | 917 | #index = tini.tm_hour*12+tini.tm_min/5 |
|
|
918 | 918 | ''' |
|
|
919 | 919 | #REVISAR |
|
|
920 | 920 | ''' |
|
|
921 | 921 | # jspc = jspc/self.nFFTPoints/self.normFactor |
|
|
922 | 922 | # jcspc = jcspc/self.nFFTPoints/self.normFactor |
|
|
923 | 923 | |
|
|
924 | 924 | |
|
|
925 | 925 | |
|
|
926 | 926 | tmp_spectra,tmp_cspectra = self.cleanRayleigh(dataOut,jspc,jcspc,self.factor_stdv) |
|
|
927 | 927 | dataOut.data_spc = tmp_spectra |
|
|
928 | 928 | dataOut.data_cspc = tmp_cspectra |
|
|
929 | 929 | |
|
|
930 | 930 | #dataOut.data_spc,dataOut.data_cspc = self.cleanRayleigh(dataOut,jspc,jcspc,self.factor_stdv) |
|
|
931 | 931 | |
|
|
932 | 932 | dataOut.data_dc = self.buffer3 |
|
|
933 | 933 | dataOut.nIncohInt *= self.nIntProfiles |
|
|
934 | 934 | dataOut.max_nIncohInt = self.nIntProfiles |
|
|
935 | 935 | dataOut.utctime = self.currentTime #tiempo promediado |
|
|
936 | 936 | #print("Time: ",time.localtime(dataOut.utctime)) |
|
|
937 | 937 | # dataOut.data_spc = sat_spectra |
|
|
938 | 938 | # dataOut.data_cspc = sat_cspectra |
|
|
939 | 939 | self.buffer = 0 |
|
|
940 | 940 | self.buffer2 = 0 |
|
|
941 | 941 | self.buffer3 = 0 |
|
|
942 | 942 | |
|
|
943 | 943 | return dataOut |
|
|
944 | 944 | |
|
|
945 | 945 | def cleanRayleigh(self,dataOut,spectra,cspectra,factor_stdv): |
|
|
946 | 946 | print("OP cleanRayleigh") |
|
|
947 | 947 | #import matplotlib.pyplot as plt |
|
|
948 | 948 | #for k in range(149): |
|
|
949 | 949 | #channelsProcssd = [] |
|
|
950 | 950 | #channelA_ok = False |
|
|
951 | 951 | #rfunc = cspectra.copy() #self.bloques |
|
|
952 | 952 | rfunc = spectra.copy() |
|
|
953 | 953 | #rfunc = cspectra |
|
|
954 | 954 | #val_spc = spectra*0.0 #self.bloque0*0.0 |
|
|
955 | 955 | #val_cspc = cspectra*0.0 #self.bloques*0.0 |
|
|
956 | 956 | #in_sat_spectra = spectra.copy() #self.bloque0 |
|
|
957 | 957 | #in_sat_cspectra = cspectra.copy() #self.bloques |
|
|
958 | 958 | |
|
|
959 | 959 | |
|
|
960 | 960 | ###ONLY FOR TEST: |
|
|
961 | 961 | raxs = math.ceil(math.sqrt(self.nPairs)) |
|
|
962 | 962 | if raxs == 0: |
|
|
963 | 963 | raxs = 1 |
|
|
964 | 964 | caxs = math.ceil(self.nPairs/raxs) |
|
|
965 | 965 | if self.nPairs <4: |
|
|
966 | 966 | raxs = 2 |
|
|
967 | 967 | caxs = 2 |
|
|
968 | 968 | #print(raxs, caxs) |
|
|
969 | 969 | fft_rev = 14 #nFFT to plot |
|
|
970 | 970 | hei_rev = ((self.heights >= 550) & (self.heights <= 551)).nonzero() #hei to plot |
|
|
971 | 971 | hei_rev = hei_rev[0] |
|
|
972 | 972 | #print(hei_rev) |
|
|
973 | 973 | |
|
|
974 | 974 | #print numpy.absolute(rfunc[:,0,0,14]) |
|
|
975 | 975 | |
|
|
976 | 976 | gauss_fit, covariance = None, None |
|
|
977 | 977 | for ih in range(self.minAltInd,self.maxAltInd): |
|
|
978 | 978 | for ifreq in range(self.nFFTPoints): |
|
|
979 | 979 | ''' |
|
|
980 | 980 | ###ONLY FOR TEST: |
|
|
981 | 981 | if ifreq ==fft_rev and ih==hei_rev: #TO VIEW A SIGNLE FREQUENCY |
|
|
982 | 982 | fig, axs = plt.subplots(raxs, caxs) |
|
|
983 | 983 | fig2, axs2 = plt.subplots(raxs, caxs) |
|
|
984 | 984 | col_ax = 0 |
|
|
985 | 985 | row_ax = 0 |
|
|
986 | 986 | ''' |
|
|
987 | 987 | #print(self.nPairs) |
|
|
988 | 988 | for ii in range(self.nChan): #PARES DE CANALES SELF y CROSS |
|
|
989 | 989 | # if self.crosspairs[ii][1]-self.crosspairs[ii][0] > 1: # APLICAR SOLO EN PARES CONTIGUOS |
|
|
990 | 990 | # continue |
|
|
991 | 991 | # if not self.crosspairs[ii][0] in channelsProcssd: |
|
|
992 | 992 | # channelA_ok = True |
|
|
993 | 993 | #print("pair: ",self.crosspairs[ii]) |
|
|
994 | 994 | ''' |
|
|
995 | 995 | ###ONLY FOR TEST: |
|
|
996 | 996 | if (col_ax%caxs==0 and col_ax!=0 and self.nPairs !=1): |
|
|
997 | 997 | col_ax = 0 |
|
|
998 | 998 | row_ax += 1 |
|
|
999 | 999 | ''' |
|
|
1000 | 1000 | func2clean = 10*numpy.log10(numpy.absolute(rfunc[:,ii,ifreq,ih])) #Potencia? |
|
|
1001 | 1001 | #print(func2clean.shape) |
|
|
1002 | 1002 | val = (numpy.isfinite(func2clean)==True).nonzero() |
|
|
1003 | 1003 | |
|
|
1004 | 1004 | if len(val)>0: #limitador |
|
|
1005 | 1005 | min_val = numpy.around(numpy.amin(func2clean)-2) #> (-40) |
|
|
1006 | 1006 | if min_val <= -40 : |
|
|
1007 | 1007 | min_val = -40 |
|
|
1008 | 1008 | max_val = numpy.around(numpy.amax(func2clean)+2) #< 200 |
|
|
1009 | 1009 | if max_val >= 200 : |
|
|
1010 | 1010 | max_val = 200 |
|
|
1011 | 1011 | #print min_val, max_val |
|
|
1012 | 1012 | step = 1 |
|
|
1013 | 1013 | #print("Getting bins and the histogram") |
|
|
1014 | 1014 | x_dist = min_val + numpy.arange(1 + ((max_val-(min_val))/step))*step |
|
|
1015 | 1015 | y_dist,binstep = numpy.histogram(func2clean,bins=range(int(min_val),int(max_val+2),step)) |
|
|
1016 | 1016 | #print(len(y_dist),len(binstep[:-1])) |
|
|
1017 | 1017 | #print(row_ax,col_ax, " ..") |
|
|
1018 | 1018 | #print(self.pairsArray[ii][0],self.pairsArray[ii][1]) |
|
|
1019 | 1019 | mean = numpy.sum(x_dist * y_dist) / numpy.sum(y_dist) |
|
|
1020 | 1020 | sigma = numpy.sqrt(numpy.sum(y_dist * (x_dist - mean)**2) / numpy.sum(y_dist)) |
|
|
1021 | 1021 | parg = [numpy.amax(y_dist),mean,sigma] |
|
|
1022 | 1022 | |
|
|
1023 | 1023 | newY = None |
|
|
1024 | 1024 | |
|
|
1025 | 1025 | try : |
|
|
1026 | 1026 | gauss_fit, covariance = curve_fit(fit_func, x_dist, y_dist,p0=parg) |
|
|
1027 | 1027 | mode = gauss_fit[1] |
|
|
1028 | 1028 | stdv = gauss_fit[2] |
|
|
1029 | 1029 | #print(" FIT OK",gauss_fit) |
|
|
1030 | 1030 | ''' |
|
|
1031 | 1031 | ###ONLY FOR TEST: |
|
|
1032 | 1032 | if ifreq ==fft_rev and ih==hei_rev: #TO VIEW A SIGNLE FREQUENCY |
|
|
1033 | 1033 | newY = fit_func(x_dist,gauss_fit[0],gauss_fit[1],gauss_fit[2]) |
|
|
1034 | 1034 | axs[row_ax,col_ax].plot(binstep[:-1],y_dist,color='green') |
|
|
1035 | 1035 | axs[row_ax,col_ax].plot(binstep[:-1],newY,color='red') |
|
|
1036 | 1036 | axs[row_ax,col_ax].set_title("CH "+str(self.channels[ii])) |
|
|
1037 | 1037 | ''' |
|
|
1038 | 1038 | except: |
|
|
1039 | 1039 | mode = mean |
|
|
1040 | 1040 | stdv = sigma |
|
|
1041 | 1041 | #print("FIT FAIL") |
|
|
1042 | 1042 | #continue |
|
|
1043 | 1043 | |
|
|
1044 | 1044 | |
|
|
1045 | 1045 | #print(mode,stdv) |
|
|
1046 | 1046 | #Removing echoes greater than mode + std_factor*stdv |
|
|
1047 | 1047 | noval = (abs(func2clean - mode)>=(factor_stdv*stdv)).nonzero() |
|
|
1048 | 1048 | #noval tiene los indices que se van a remover |
|
|
1049 | 1049 | #print("Chan ",ii," novals: ",len(noval[0])) |
|
|
1050 | 1050 | if len(noval[0]) > 0: #forma de array (N,) es igual a longitud (N) |
|
|
1051 | 1051 | novall = ((func2clean - mode) >= (factor_stdv*stdv)).nonzero() |
|
|
1052 | 1052 | #print(novall) |
|
|
1053 | 1053 | #print(" ",self.pairsArray[ii]) |
|
|
1054 | 1054 | #cross_pairs = self.pairsArray[ii] |
|
|
1055 | 1055 | #Getting coherent echoes which are removed. |
|
|
1056 | 1056 | # if len(novall[0]) > 0: |
|
|
1057 | 1057 | # |
|
|
1058 | 1058 | # val_spc[novall[0],cross_pairs[0],ifreq,ih] = 1 |
|
|
1059 | 1059 | # val_spc[novall[0],cross_pairs[1],ifreq,ih] = 1 |
|
|
1060 | 1060 | # val_cspc[novall[0],ii,ifreq,ih] = 1 |
|
|
1061 | 1061 | #print("OUT NOVALL 1") |
|
|
1062 | 1062 | try: |
|
|
1063 | 1063 | pair = (self.channels[ii],self.channels[ii + 1]) |
|
|
1064 | 1064 | except: |
|
|
1065 | 1065 | pair = (99,99) |
|
|
1066 | 1066 | #print("par ", pair) |
|
|
1067 | 1067 | if ( pair in self.crosspairs): |
|
|
1068 | 1068 | q = self.crosspairs.index(pair) |
|
|
1069 | 1069 | #print("estΓ‘ aqui: ", q, (ii,ii + 1)) |
|
|
1070 | 1070 | new_a = numpy.delete(cspectra[:,q,ifreq,ih], noval[0]) |
|
|
1071 | 1071 | cspectra[noval,q,ifreq,ih] = numpy.mean(new_a) #mean CrossSpectra |
|
|
1072 | 1072 | |
|
|
1073 | 1073 | #if channelA_ok: |
|
|
1074 | 1074 | #chA = self.channels.index(cross_pairs[0]) |
|
|
1075 | 1075 | new_b = numpy.delete(spectra[:,ii,ifreq,ih], noval[0]) |
|
|
1076 | 1076 | spectra[noval,ii,ifreq,ih] = numpy.mean(new_b) #mean Spectra Pair A |
|
|
1077 | 1077 | #channelA_ok = False |
|
|
1078 | 1078 | |
|
|
1079 | 1079 | # chB = self.channels.index(cross_pairs[1]) |
|
|
1080 | 1080 | # new_c = numpy.delete(spectra[:,chB,ifreq,ih], noval[0]) |
|
|
1081 | 1081 | # spectra[noval,chB,ifreq,ih] = numpy.mean(new_c) #mean Spectra Pair B |
|
|
1082 | 1082 | # |
|
|
1083 | 1083 | # channelsProcssd.append(self.crosspairs[ii][0]) # save channel A |
|
|
1084 | 1084 | # channelsProcssd.append(self.crosspairs[ii][1]) # save channel B |
|
|
1085 | 1085 | ''' |
|
|
1086 | 1086 | ###ONLY FOR TEST: |
|
|
1087 | 1087 | if ifreq ==fft_rev and ih==hei_rev: #TO VIEW A SIGNLE FREQUENCY |
|
|
1088 | 1088 | func2clean = 10*numpy.log10(numpy.absolute(spectra[:,ii,ifreq,ih])) |
|
|
1089 | 1089 | y_dist,binstep = numpy.histogram(func2clean,bins=range(int(min_val),int(max_val+2),step)) |
|
|
1090 | 1090 | axs2[row_ax,col_ax].plot(binstep[:-1],newY,color='red') |
|
|
1091 | 1091 | axs2[row_ax,col_ax].plot(binstep[:-1],y_dist,color='green') |
|
|
1092 | 1092 | axs2[row_ax,col_ax].set_title("CH "+str(self.channels[ii])) |
|
|
1093 | 1093 | ''' |
|
|
1094 | 1094 | ''' |
|
|
1095 | 1095 | ###ONLY FOR TEST: |
|
|
1096 | 1096 | col_ax += 1 #contador de ploteo columnas |
|
|
1097 | 1097 | ##print(col_ax) |
|
|
1098 | 1098 | ###ONLY FOR TEST: |
|
|
1099 | 1099 | if ifreq ==fft_rev and ih==hei_rev: #TO VIEW A SIGNLE FREQUENCY |
|
|
1100 | 1100 | title = str(dataOut.datatime)+" nFFT: "+str(ifreq)+" Alt: "+str(self.heights[ih])+ " km" |
|
|
1101 | 1101 | title2 = str(dataOut.datatime)+" nFFT: "+str(ifreq)+" Alt: "+str(self.heights[ih])+ " km CLEANED" |
|
|
1102 | 1102 | fig.suptitle(title) |
|
|
1103 | 1103 | fig2.suptitle(title2) |
|
|
1104 | 1104 | plt.show() |
|
|
1105 | 1105 | ''' |
|
|
1106 | 1106 | ################################################################################################## |
|
|
1107 | 1107 | |
|
|
1108 | 1108 | #print("Getting average of the spectra and cross-spectra from incoherent echoes.") |
|
|
1109 | 1109 | out_spectra = numpy.zeros([self.nChan,self.nFFTPoints,self.nHeights], dtype=float) #+numpy.nan |
|
|
1110 | 1110 | out_cspectra = numpy.zeros([self.nPairs,self.nFFTPoints,self.nHeights], dtype=complex) #+numpy.nan |
|
|
1111 | 1111 | for ih in range(self.nHeights): |
|
|
1112 | 1112 | for ifreq in range(self.nFFTPoints): |
|
|
1113 | 1113 | for ich in range(self.nChan): |
|
|
1114 | 1114 | tmp = spectra[:,ich,ifreq,ih] |
|
|
1115 | 1115 | valid = (numpy.isfinite(tmp[:])==True).nonzero() |
|
|
1116 | 1116 | |
|
|
1117 | 1117 | if len(valid[0]) >0 : |
|
|
1118 | 1118 | out_spectra[ich,ifreq,ih] = numpy.nansum(tmp)#/len(valid[0]) |
|
|
1119 | 1119 | |
|
|
1120 | 1120 | for icr in range(self.nPairs): |
|
|
1121 | 1121 | tmp = numpy.squeeze(cspectra[:,icr,ifreq,ih]) |
|
|
1122 | 1122 | valid = (numpy.isfinite(tmp)==True).nonzero() |
|
|
1123 | 1123 | if len(valid[0]) > 0: |
|
|
1124 | 1124 | out_cspectra[icr,ifreq,ih] = numpy.nansum(tmp)#/len(valid[0]) |
|
|
1125 | 1125 | |
|
|
1126 | 1126 | return out_spectra, out_cspectra |
|
|
1127 | 1127 | |
|
|
1128 | 1128 | def REM_ISOLATED_POINTS(self,array,rth): |
|
|
1129 | 1129 | # import matplotlib.pyplot as plt |
|
|
1130 | 1130 | if rth == None : |
|
|
1131 | 1131 | rth = 4 |
|
|
1132 | 1132 | #print("REM ISO") |
|
|
1133 | 1133 | num_prof = len(array[0,:,0]) |
|
|
1134 | 1134 | num_hei = len(array[0,0,:]) |
|
|
1135 | 1135 | n2d = len(array[:,0,0]) |
|
|
1136 | 1136 | |
|
|
1137 | 1137 | for ii in range(n2d) : |
|
|
1138 | 1138 | #print ii,n2d |
|
|
1139 | 1139 | tmp = array[ii,:,:] |
|
|
1140 | 1140 | #print tmp.shape, array[ii,101,:],array[ii,102,:] |
|
|
1141 | 1141 | |
|
|
1142 | 1142 | # fig = plt.figure(figsize=(6,5)) |
|
|
1143 | 1143 | # left, bottom, width, height = 0.1, 0.1, 0.8, 0.8 |
|
|
1144 | 1144 | # ax = fig.add_axes([left, bottom, width, height]) |
|
|
1145 | 1145 | # x = range(num_prof) |
|
|
1146 | 1146 | # y = range(num_hei) |
|
|
1147 | 1147 | # cp = ax.contour(y,x,tmp) |
|
|
1148 | 1148 | # ax.clabel(cp, inline=True,fontsize=10) |
|
|
1149 | 1149 | # plt.show() |
|
|
1150 | 1150 | |
|
|
1151 | 1151 | #indxs = WHERE(FINITE(tmp) AND tmp GT 0,cindxs) |
|
|
1152 | 1152 | tmp = numpy.reshape(tmp,num_prof*num_hei) |
|
|
1153 | 1153 | indxs1 = (numpy.isfinite(tmp)==True).nonzero() |
|
|
1154 | 1154 | indxs2 = (tmp > 0).nonzero() |
|
|
1155 | 1155 | |
|
|
1156 | 1156 | indxs1 = (indxs1[0]) |
|
|
1157 | 1157 | indxs2 = indxs2[0] |
|
|
1158 | 1158 | #indxs1 = numpy.array(indxs1[0]) |
|
|
1159 | 1159 | #indxs2 = numpy.array(indxs2[0]) |
|
|
1160 | 1160 | indxs = None |
|
|
1161 | 1161 | #print indxs1 , indxs2 |
|
|
1162 | 1162 | for iv in range(len(indxs2)): |
|
|
1163 | 1163 | indv = numpy.array((indxs1 == indxs2[iv]).nonzero()) |
|
|
1164 | 1164 | #print len(indxs2), indv |
|
|
1165 | 1165 | if len(indv[0]) > 0 : |
|
|
1166 | 1166 | indxs = numpy.concatenate((indxs,indxs2[iv]), axis=None) |
|
|
1167 | 1167 | # print indxs |
|
|
1168 | 1168 | indxs = indxs[1:] |
|
|
1169 | 1169 | #print(indxs, len(indxs)) |
|
|
1170 | 1170 | if len(indxs) < 4 : |
|
|
1171 | 1171 | array[ii,:,:] = 0. |
|
|
1172 | 1172 | return |
|
|
1173 | 1173 | |
|
|
1174 | 1174 | xpos = numpy.mod(indxs ,num_hei) |
|
|
1175 | 1175 | ypos = (indxs / num_hei) |
|
|
1176 | 1176 | sx = numpy.argsort(xpos) # Ordering respect to "x" (time) |
|
|
1177 | 1177 | #print sx |
|
|
1178 | 1178 | xpos = xpos[sx] |
|
|
1179 | 1179 | ypos = ypos[sx] |
|
|
1180 | 1180 | |
|
|
1181 | 1181 | # *********************************** Cleaning isolated points ********************************** |
|
|
1182 | 1182 | ic = 0 |
|
|
1183 | 1183 | while True : |
|
|
1184 | 1184 | r = numpy.sqrt(list(numpy.power((xpos[ic]-xpos),2)+ numpy.power((ypos[ic]-ypos),2))) |
|
|
1185 | 1185 | #no_coh = WHERE(FINITE(r) AND (r LE rth),cno_coh) |
|
|
1186 | 1186 | #plt.plot(r) |
|
|
1187 | 1187 | #plt.show() |
|
|
1188 | 1188 | no_coh1 = (numpy.isfinite(r)==True).nonzero() |
|
|
1189 | 1189 | no_coh2 = (r <= rth).nonzero() |
|
|
1190 | 1190 | #print r, no_coh1, no_coh2 |
|
|
1191 | 1191 | no_coh1 = numpy.array(no_coh1[0]) |
|
|
1192 | 1192 | no_coh2 = numpy.array(no_coh2[0]) |
|
|
1193 | 1193 | no_coh = None |
|
|
1194 | 1194 | #print valid1 , valid2 |
|
|
1195 | 1195 | for iv in range(len(no_coh2)): |
|
|
1196 | 1196 | indv = numpy.array((no_coh1 == no_coh2[iv]).nonzero()) |
|
|
1197 | 1197 | if len(indv[0]) > 0 : |
|
|
1198 | 1198 | no_coh = numpy.concatenate((no_coh,no_coh2[iv]), axis=None) |
|
|
1199 | 1199 | no_coh = no_coh[1:] |
|
|
1200 | 1200 | #print len(no_coh), no_coh |
|
|
1201 | 1201 | if len(no_coh) < 4 : |
|
|
1202 | 1202 | #print xpos[ic], ypos[ic], ic |
|
|
1203 | 1203 | # plt.plot(r) |
|
|
1204 | 1204 | # plt.show() |
|
|
1205 | 1205 | xpos[ic] = numpy.nan |
|
|
1206 | 1206 | ypos[ic] = numpy.nan |
|
|
1207 | 1207 | |
|
|
1208 | 1208 | ic = ic + 1 |
|
|
1209 | 1209 | if (ic == len(indxs)) : |
|
|
1210 | 1210 | break |
|
|
1211 | 1211 | #print( xpos, ypos) |
|
|
1212 | 1212 | |
|
|
1213 | 1213 | indxs = (numpy.isfinite(list(xpos))==True).nonzero() |
|
|
1214 | 1214 | #print indxs[0] |
|
|
1215 | 1215 | if len(indxs[0]) < 4 : |
|
|
1216 | 1216 | array[ii,:,:] = 0. |
|
|
1217 | 1217 | return |
|
|
1218 | 1218 | |
|
|
1219 | 1219 | xpos = xpos[indxs[0]] |
|
|
1220 | 1220 | ypos = ypos[indxs[0]] |
|
|
1221 | 1221 | for i in range(0,len(ypos)): |
|
|
1222 | 1222 | ypos[i]=int(ypos[i]) |
|
|
1223 | 1223 | junk = tmp |
|
|
1224 | 1224 | tmp = junk*0.0 |
|
|
1225 | 1225 | |
|
|
1226 | 1226 | tmp[list(xpos + (ypos*num_hei))] = junk[list(xpos + (ypos*num_hei))] |
|
|
1227 | 1227 | array[ii,:,:] = numpy.reshape(tmp,(num_prof,num_hei)) |
|
|
1228 | 1228 | |
|
|
1229 | 1229 | #print array.shape |
|
|
1230 | 1230 | #tmp = numpy.reshape(tmp,(num_prof,num_hei)) |
|
|
1231 | 1231 | #print tmp.shape |
|
|
1232 | 1232 | |
|
|
1233 | 1233 | # fig = plt.figure(figsize=(6,5)) |
|
|
1234 | 1234 | # left, bottom, width, height = 0.1, 0.1, 0.8, 0.8 |
|
|
1235 | 1235 | # ax = fig.add_axes([left, bottom, width, height]) |
|
|
1236 | 1236 | # x = range(num_prof) |
|
|
1237 | 1237 | # y = range(num_hei) |
|
|
1238 | 1238 | # cp = ax.contour(y,x,array[ii,:,:]) |
|
|
1239 | 1239 | # ax.clabel(cp, inline=True,fontsize=10) |
|
|
1240 | 1240 | # plt.show() |
|
|
1241 | 1241 | return array |
|
|
1242 | 1242 | |
|
|
1243 | 1243 | |
|
|
1244 | 1244 | class IntegrationFaradaySpectra(Operation): |
|
|
1245 | 1245 | |
|
|
1246 | 1246 | __profIndex = 0 |
|
|
1247 | 1247 | __withOverapping = False |
|
|
1248 | 1248 | |
|
|
1249 | 1249 | __byTime = False |
|
|
1250 | 1250 | __initime = None |
|
|
1251 | 1251 | __lastdatatime = None |
|
|
1252 | 1252 | __integrationtime = None |
|
|
1253 | 1253 | |
|
|
1254 | 1254 | __buffer_spc = None |
|
|
1255 | 1255 | __buffer_cspc = None |
|
|
1256 | 1256 | __buffer_dc = None |
|
|
1257 | 1257 | |
|
|
1258 | 1258 | __dataReady = False |
|
|
1259 | 1259 | |
|
|
1260 | 1260 | __timeInterval = None |
|
|
1261 | 1261 | n_ints = None #matriz de numero de integracions (CH,HEI) |
|
|
1262 | 1262 | n = None |
|
|
1263 | 1263 | minHei_ind = None |
|
|
1264 | 1264 | maxHei_ind = None |
|
|
1265 | 1265 | navg = 1.0 |
|
|
1266 | 1266 | factor = 0.0 |
|
|
1267 | 1267 | dataoutliers = None # (CHANNELS, HEIGHTS) |
|
|
1268 | 1268 | |
|
|
1269 | 1269 | def __init__(self): |
|
|
1270 | 1270 | |
|
|
1271 | 1271 | Operation.__init__(self) |
|
|
1272 | 1272 | |
|
|
1273 | 1273 | def setup(self, dataOut,n=None, timeInterval=None, overlapping=False, DPL=None, minHei=None, maxHei=None, avg=1,factor=0.75): |
|
|
1274 | 1274 | """ |
|
|
1275 | 1275 | Set the parameters of the integration class. |
|
|
1276 | 1276 | |
|
|
1277 | 1277 | Inputs: |
|
|
1278 | 1278 | |
|
|
1279 | 1279 | n : Number of coherent integrations |
|
|
1280 | 1280 | timeInterval : Time of integration. If the parameter "n" is selected this one does not work |
|
|
1281 | 1281 | overlapping : |
|
|
1282 | 1282 | |
|
|
1283 | 1283 | """ |
|
|
1284 | 1284 | |
|
|
1285 | 1285 | self.__initime = None |
|
|
1286 | 1286 | self.__lastdatatime = 0 |
|
|
1287 | 1287 | |
|
|
1288 | 1288 | self.__buffer_spc = [] |
|
|
1289 | 1289 | self.__buffer_cspc = [] |
|
|
1290 | 1290 | self.__buffer_dc = 0 |
|
|
1291 | 1291 | |
|
|
1292 | 1292 | self.__profIndex = 0 |
|
|
1293 | 1293 | self.__dataReady = False |
|
|
1294 | 1294 | self.__byTime = False |
|
|
1295 | 1295 | |
|
|
1296 | 1296 | self.factor = factor |
|
|
1297 | 1297 | self.navg = avg |
|
|
1298 | 1298 | #self.ByLags = dataOut.ByLags ###REDEFINIR |
|
|
1299 | 1299 | self.ByLags = False |
|
|
1300 | 1300 | self.maxProfilesInt = 1 |
|
|
1301 | 1301 | |
|
|
1302 | 1302 | if DPL != None: |
|
|
1303 | 1303 | self.DPL=DPL |
|
|
1304 | 1304 | else: |
|
|
1305 | 1305 | #self.DPL=dataOut.DPL ###REDEFINIR |
|
|
1306 | 1306 | self.DPL=0 |
|
|
1307 | 1307 | |
|
|
1308 | 1308 | if n is None and timeInterval is None: |
|
|
1309 | 1309 | raise ValueError("n or timeInterval should be specified ...") |
|
|
1310 | 1310 | |
|
|
1311 | 1311 | if n is not None: |
|
|
1312 | 1312 | self.n = int(n) |
|
|
1313 | 1313 | else: |
|
|
1314 | 1314 | self.__integrationtime = int(timeInterval) |
|
|
1315 | 1315 | self.n = None |
|
|
1316 | 1316 | self.__byTime = True |
|
|
1317 | 1317 | |
|
|
1318 | 1318 | if minHei == None: |
|
|
1319 | 1319 | minHei = self.dataOut.heightList[0] |
|
|
1320 | 1320 | |
|
|
1321 | 1321 | if maxHei == None: |
|
|
1322 | 1322 | maxHei = self.dataOut.heightList[-1] |
|
|
1323 | 1323 | |
|
|
1324 | 1324 | if (minHei < self.dataOut.heightList[0]) or (minHei > maxHei): |
|
|
1325 | 1325 | print('minHei: %.2f is out of the heights range' % (minHei)) |
|
|
1326 | 1326 | print('minHei is setting to %.2f' % (self.dataOut.heightList[0])) |
|
|
1327 | 1327 | minHei = self.dataOut.heightList[0] |
|
|
1328 | 1328 | |
|
|
1329 | 1329 | if (maxHei > self.dataOut.heightList[-1]) or (maxHei < minHei): |
|
|
1330 | 1330 | print('maxHei: %.2f is out of the heights range' % (maxHei)) |
|
|
1331 | 1331 | print('maxHei is setting to %.2f' % (self.dataOut.heightList[-1])) |
|
|
1332 | 1332 | maxHei = self.dataOut.heightList[-1] |
|
|
1333 | 1333 | |
|
|
1334 | 1334 | ind_list1 = numpy.where(self.dataOut.heightList >= minHei) |
|
|
1335 | 1335 | ind_list2 = numpy.where(self.dataOut.heightList <= maxHei) |
|
|
1336 | 1336 | self.minHei_ind = ind_list1[0][0] |
|
|
1337 | 1337 | self.maxHei_ind = ind_list2[0][-1] |
|
|
1338 | 1338 | |
|
|
1339 | 1339 | def putData(self, data_spc, data_cspc, data_dc): |
|
|
1340 | 1340 | """ |
|
|
1341 | 1341 | Add a profile to the __buffer_spc and increase in one the __profileIndex |
|
|
1342 | 1342 | |
|
|
1343 | 1343 | """ |
|
|
1344 | 1344 | |
|
|
1345 | 1345 | self.__buffer_spc.append(data_spc) |
|
|
1346 | 1346 | |
|
|
1347 | 1347 | if data_cspc is None: |
|
|
1348 | 1348 | self.__buffer_cspc = None |
|
|
1349 | 1349 | else: |
|
|
1350 | 1350 | self.__buffer_cspc.append(data_cspc) |
|
|
1351 | 1351 | |
|
|
1352 | 1352 | if data_dc is None: |
|
|
1353 | 1353 | self.__buffer_dc = None |
|
|
1354 | 1354 | else: |
|
|
1355 | 1355 | self.__buffer_dc += data_dc |
|
|
1356 | 1356 | |
|
|
1357 | 1357 | self.__profIndex += 1 |
|
|
1358 | 1358 | |
|
|
1359 | 1359 | return |
|
|
1360 | 1360 | |
|
|
1361 | 1361 | def hildebrand_sekhon_Integration(self,sortdata,navg, factor): |
|
|
1362 | 1362 | #data debe estar ordenado |
|
|
1363 | 1363 | #sortdata = numpy.sort(data, axis=None) |
|
|
1364 | 1364 | #sortID=data.argsort() |
|
|
1365 | 1365 | lenOfData = len(sortdata) |
|
|
1366 | 1366 | nums_min = lenOfData*factor |
|
|
1367 | 1367 | if nums_min <= 5: |
|
|
1368 | 1368 | nums_min = 5 |
|
|
1369 | 1369 | sump = 0. |
|
|
1370 | 1370 | sumq = 0. |
|
|
1371 | 1371 | j = 0 |
|
|
1372 | 1372 | cont = 1 |
|
|
1373 | 1373 | while((cont == 1)and(j < lenOfData)): |
|
|
1374 | 1374 | sump += sortdata[j] |
|
|
1375 | 1375 | sumq += sortdata[j]**2 |
|
|
1376 | 1376 | if j > nums_min: |
|
|
1377 | 1377 | rtest = float(j)/(j-1) + 1.0/navg |
|
|
1378 | 1378 | if ((sumq*j) > (rtest*sump**2)): |
|
|
1379 | 1379 | j = j - 1 |
|
|
1380 | 1380 | sump = sump - sortdata[j] |
|
|
1381 | 1381 | sumq = sumq - sortdata[j]**2 |
|
|
1382 | 1382 | cont = 0 |
|
|
1383 | 1383 | j += 1 |
|
|
1384 | 1384 | #lnoise = sump / j |
|
|
1385 | 1385 | #print("H S done") |
|
|
1386 | 1386 | #return j,sortID |
|
|
1387 | 1387 | return j |
|
|
1388 | 1388 | |
|
|
1389 | 1389 | |
|
|
1390 | 1390 | def pushData(self): |
|
|
1391 | 1391 | """ |
|
|
1392 | 1392 | Return the sum of the last profiles and the profiles used in the sum. |
|
|
1393 | 1393 | |
|
|
1394 | 1394 | Affected: |
|
|
1395 | 1395 | |
|
|
1396 | 1396 | self.__profileIndex |
|
|
1397 | 1397 | |
|
|
1398 | 1398 | """ |
|
|
1399 | 1399 | bufferH=None |
|
|
1400 | 1400 | buffer=None |
|
|
1401 | 1401 | buffer1=None |
|
|
1402 | 1402 | buffer_cspc=None |
|
|
1403 | 1403 | self.__buffer_spc=numpy.array(self.__buffer_spc) |
|
|
1404 | 1404 | try: |
|
|
1405 | 1405 | self.__buffer_cspc=numpy.array(self.__buffer_cspc) |
|
|
1406 | 1406 | except : |
|
|
1407 | 1407 | #print("No cpsc",e) |
|
|
1408 | 1408 | pass |
|
|
1409 | 1409 | #print("FREQ_DC",self.__buffer_spc.shape,self.__buffer_cspc.shape) |
|
|
1410 | 1410 | |
|
|
1411 | 1411 | freq_dc = int(self.__buffer_spc.shape[2] / 2) |
|
|
1412 | 1412 | #print("FREQ_DC",freq_dc,self.__buffer_spc.shape,self.nHeights) |
|
|
1413 | 1413 | |
|
|
1414 | 1414 | self.dataOutliers = numpy.zeros((self.nChannels,self.nHeights)) # --> almacen de outliers |
|
|
1415 | 1415 | |
|
|
1416 | 1416 | for k in range(self.minHei_ind,self.maxHei_ind): |
|
|
1417 | 1417 | try: |
|
|
1418 | 1418 | buffer_cspc=numpy.copy(self.__buffer_cspc[:,:,:,k]) |
|
|
1419 | 1419 | except: |
|
|
1420 | 1420 | #print("No cpsc",e) |
|
|
1421 | 1421 | pass |
|
|
1422 | 1422 | outliers_IDs_cspc=[] |
|
|
1423 | 1423 | cspc_outliers_exist=False |
|
|
1424 | 1424 | for i in range(self.nChannels):#dataOut.nChannels): |
|
|
1425 | 1425 | |
|
|
1426 | 1426 | buffer1=numpy.copy(self.__buffer_spc[:,i,:,k]) |
|
|
1427 | 1427 | indexes=[] |
|
|
1428 | 1428 | #sortIDs=[] |
|
|
1429 | 1429 | outliers_IDs=[] |
|
|
1430 | 1430 | |
|
|
1431 | 1431 | for j in range(self.nProfiles): #frecuencias en el tiempo |
|
|
1432 | 1432 | # if i==0 and j==freq_dc: #NOT CONSIDERING DC PROFILE AT CHANNEL 0 |
|
|
1433 | 1433 | # continue |
|
|
1434 | 1434 | # if i==1 and j==0: #NOT CONSIDERING DC PROFILE AT CHANNEL 1 |
|
|
1435 | 1435 | # continue |
|
|
1436 | 1436 | buffer=buffer1[:,j] |
|
|
1437 | 1437 | sortdata = numpy.sort(buffer, axis=None) |
|
|
1438 | 1438 | |
|
|
1439 | 1439 | sortID=buffer.argsort() |
|
|
1440 | 1440 | index = _noise.hildebrand_sekhon2(sortdata,self.navg) |
|
|
1441 | 1441 | |
|
|
1442 | 1442 | #index,sortID=self.hildebrand_sekhon_Integration(buffer,1,self.factor) |
|
|
1443 | 1443 | |
|
|
1444 | 1444 | # fig,ax = plt.subplots() |
|
|
1445 | 1445 | # ax.set_title(str(k)+" "+str(j)) |
|
|
1446 | 1446 | # x=range(len(sortdata)) |
|
|
1447 | 1447 | # ax.scatter(x,sortdata) |
|
|
1448 | 1448 | # ax.axvline(index) |
|
|
1449 | 1449 | # plt.show() |
|
|
1450 | 1450 | |
|
|
1451 | 1451 | indexes.append(index) |
|
|
1452 | 1452 | #sortIDs.append(sortID) |
|
|
1453 | 1453 | outliers_IDs=numpy.append(outliers_IDs,sortID[index:]) |
|
|
1454 | 1454 | |
|
|
1455 | 1455 | #print("Outliers: ",outliers_IDs) |
|
|
1456 | 1456 | outliers_IDs=numpy.array(outliers_IDs) |
|
|
1457 | 1457 | outliers_IDs=outliers_IDs.ravel() |
|
|
1458 | 1458 | outliers_IDs=numpy.unique(outliers_IDs) |
|
|
1459 | 1459 | outliers_IDs=outliers_IDs.astype(numpy.dtype('int64')) |
|
|
1460 | 1460 | indexes=numpy.array(indexes) |
|
|
1461 | 1461 | indexmin=numpy.min(indexes) |
|
|
1462 | 1462 | |
|
|
1463 | 1463 | |
|
|
1464 | 1464 | #print(indexmin,buffer1.shape[0], k) |
|
|
1465 | 1465 | |
|
|
1466 | 1466 | # fig,ax = plt.subplots() |
|
|
1467 | 1467 | # ax.plot(sortdata) |
|
|
1468 | 1468 | # ax2 = ax.twinx() |
|
|
1469 | 1469 | # x=range(len(indexes)) |
|
|
1470 | 1470 | # #plt.scatter(x,indexes) |
|
|
1471 | 1471 | # ax2.scatter(x,indexes) |
|
|
1472 | 1472 | # plt.show() |
|
|
1473 | 1473 | |
|
|
1474 | 1474 | if indexmin != buffer1.shape[0]: |
|
|
1475 | 1475 | if self.nChannels > 1: |
|
|
1476 | 1476 | cspc_outliers_exist= True |
|
|
1477 | 1477 | |
|
|
1478 | 1478 | lt=outliers_IDs |
|
|
1479 | 1479 | #avg=numpy.mean(buffer1[[t for t in range(buffer1.shape[0]) if t not in lt],:],axis=0) |
|
|
1480 | 1480 | |
|
|
1481 | 1481 | for p in list(outliers_IDs): |
|
|
1482 | 1482 | #buffer1[p,:]=avg |
|
|
1483 | 1483 | buffer1[p,:] = numpy.NaN |
|
|
1484 | 1484 | |
|
|
1485 | 1485 | self.dataOutliers[i,k] = len(outliers_IDs) |
|
|
1486 | 1486 | |
|
|
1487 | 1487 | self.__buffer_spc[:,i,:,k]=numpy.copy(buffer1) |
|
|
1488 | 1488 | |
|
|
1489 | 1489 | outliers_IDs_cspc=numpy.append(outliers_IDs_cspc,outliers_IDs) |
|
|
1490 | 1490 | |
|
|
1491 | 1491 | |
|
|
1492 | 1492 | |
|
|
1493 | 1493 | outliers_IDs_cspc=outliers_IDs_cspc.astype(numpy.dtype('int64')) |
|
|
1494 | 1494 | if cspc_outliers_exist : |
|
|
1495 | 1495 | |
|
|
1496 | 1496 | lt=outliers_IDs_cspc |
|
|
1497 | 1497 | |
|
|
1498 | 1498 | #avg=numpy.mean(buffer_cspc[[t for t in range(buffer_cspc.shape[0]) if t not in lt],:],axis=0) |
|
|
1499 | 1499 | for p in list(outliers_IDs_cspc): |
|
|
1500 | 1500 | #buffer_cspc[p,:]=avg |
|
|
1501 | 1501 | buffer_cspc[p,:] = numpy.NaN |
|
|
1502 | 1502 | |
|
|
1503 | 1503 | try: |
|
|
1504 | 1504 | self.__buffer_cspc[:,:,:,k]=numpy.copy(buffer_cspc) |
|
|
1505 | 1505 | except: |
|
|
1506 | 1506 | #print("No cpsc",e) |
|
|
1507 | 1507 | pass |
|
|
1508 | 1508 | #else: |
|
|
1509 | 1509 | #break |
|
|
1510 | 1510 | |
|
|
1511 | 1511 | |
|
|
1512 | 1512 | |
|
|
1513 | 1513 | nOutliers = len(outliers_IDs) |
|
|
1514 | 1514 | #print("Outliers n: ",self.dataOutliers,nOutliers) |
|
|
1515 | 1515 | buffer=None |
|
|
1516 | 1516 | bufferH=None |
|
|
1517 | 1517 | buffer1=None |
|
|
1518 | 1518 | buffer_cspc=None |
|
|
1519 | 1519 | |
|
|
1520 | 1520 | |
|
|
1521 | 1521 | buffer=None |
|
|
1522 | 1522 | |
|
|
1523 | 1523 | #data_spc = numpy.sum(self.__buffer_spc,axis=0) |
|
|
1524 | 1524 | data_spc = numpy.nansum(self.__buffer_spc,axis=0) |
|
|
1525 | 1525 | try: |
|
|
1526 | 1526 | #data_cspc = numpy.sum(self.__buffer_cspc,axis=0) |
|
|
1527 | 1527 | data_cspc = numpy.nansum(self.__buffer_cspc,axis=0) |
|
|
1528 | 1528 | except: |
|
|
1529 | 1529 | #print("No cpsc",e) |
|
|
1530 | 1530 | pass |
|
|
1531 | 1531 | |
|
|
1532 | 1532 | |
|
|
1533 | 1533 | data_dc = self.__buffer_dc |
|
|
1534 | 1534 | #(CH, HEIGH) |
|
|
1535 | 1535 | self.maxProfilesInt = self.__profIndex |
|
|
1536 | 1536 | n = self.__profIndex - self.dataOutliers # n becomes a matrix |
|
|
1537 | 1537 | |
|
|
1538 | 1538 | self.__buffer_spc = [] |
|
|
1539 | 1539 | self.__buffer_cspc = [] |
|
|
1540 | 1540 | self.__buffer_dc = 0 |
|
|
1541 | 1541 | self.__profIndex = 0 |
|
|
1542 | 1542 | |
|
|
1543 | 1543 | return data_spc, data_cspc, data_dc, n |
|
|
1544 | 1544 | |
|
|
1545 | 1545 | def byProfiles(self, *args): |
|
|
1546 | 1546 | |
|
|
1547 | 1547 | self.__dataReady = False |
|
|
1548 | 1548 | avgdata_spc = None |
|
|
1549 | 1549 | avgdata_cspc = None |
|
|
1550 | 1550 | avgdata_dc = None |
|
|
1551 | 1551 | |
|
|
1552 | 1552 | self.putData(*args) |
|
|
1553 | 1553 | |
|
|
1554 | 1554 | if self.__profIndex == self.n: |
|
|
1555 | 1555 | |
|
|
1556 | 1556 | avgdata_spc, avgdata_cspc, avgdata_dc, n = self.pushData() |
|
|
1557 | 1557 | self.n_ints = n |
|
|
1558 | 1558 | self.__dataReady = True |
|
|
1559 | 1559 | |
|
|
1560 | 1560 | return avgdata_spc, avgdata_cspc, avgdata_dc |
|
|
1561 | 1561 | |
|
|
1562 | 1562 | def byTime(self, datatime, *args): |
|
|
1563 | 1563 | |
|
|
1564 | 1564 | self.__dataReady = False |
|
|
1565 | 1565 | avgdata_spc = None |
|
|
1566 | 1566 | avgdata_cspc = None |
|
|
1567 | 1567 | avgdata_dc = None |
|
|
1568 | 1568 | |
|
|
1569 | 1569 | self.putData(*args) |
|
|
1570 | 1570 | |
|
|
1571 | 1571 | if (datatime - self.__initime) >= self.__integrationtime: |
|
|
1572 | 1572 | avgdata_spc, avgdata_cspc, avgdata_dc, n = self.pushData() |
|
|
1573 | 1573 | self.n_ints = n |
|
|
1574 | 1574 | self.__dataReady = True |
|
|
1575 | 1575 | |
|
|
1576 | 1576 | return avgdata_spc, avgdata_cspc, avgdata_dc |
|
|
1577 | 1577 | |
|
|
1578 | 1578 | def integrate(self, datatime, *args): |
|
|
1579 | 1579 | |
|
|
1580 | 1580 | if self.__profIndex == 0: |
|
|
1581 | 1581 | self.__initime = datatime |
|
|
1582 | 1582 | |
|
|
1583 | 1583 | if self.__byTime: |
|
|
1584 | 1584 | avgdata_spc, avgdata_cspc, avgdata_dc = self.byTime( |
|
|
1585 | 1585 | datatime, *args) |
|
|
1586 | 1586 | else: |
|
|
1587 | 1587 | avgdata_spc, avgdata_cspc, avgdata_dc = self.byProfiles(*args) |
|
|
1588 | 1588 | |
|
|
1589 | 1589 | if not self.__dataReady: |
|
|
1590 | 1590 | return None, None, None, None |
|
|
1591 | 1591 | |
|
|
1592 | 1592 | return self.__initime, avgdata_spc, avgdata_cspc, avgdata_dc |
|
|
1593 | 1593 | |
|
|
1594 | 1594 | def run(self, dataOut, n=None, DPL = None,timeInterval=None, overlapping=False, minHei=None, maxHei=None, avg=1, factor=0.75): |
|
|
1595 | 1595 | self.dataOut = dataOut |
|
|
1596 | 1596 | if n == 1: |
|
|
1597 | 1597 | return self.dataOut |
|
|
1598 | 1598 | |
|
|
1599 | 1599 | |
|
|
1600 | 1600 | if self.dataOut.nChannels == 1: |
|
|
1601 | 1601 | self.dataOut.data_cspc = None #si es un solo canal no vale la pena acumular DATOS |
|
|
1602 | 1602 | #print(self.dataOut.data_spc.shape, self.dataOut.data_cspc) |
|
|
1603 | 1603 | if not self.isConfig: |
|
|
1604 | 1604 | self.setup(self.dataOut, n, timeInterval, overlapping,DPL ,minHei, maxHei, avg, factor) |
|
|
1605 | 1605 | self.isConfig = True |
|
|
1606 | 1606 | |
|
|
1607 | 1607 | if not self.ByLags: |
|
|
1608 | 1608 | self.nProfiles=self.dataOut.nProfiles |
|
|
1609 | 1609 | self.nChannels=self.dataOut.nChannels |
|
|
1610 | 1610 | self.nHeights=self.dataOut.nHeights |
|
|
1611 | 1611 | avgdatatime, avgdata_spc, avgdata_cspc, avgdata_dc = self.integrate(self.dataOut.utctime, |
|
|
1612 | 1612 | self.dataOut.data_spc, |
|
|
1613 | 1613 | self.dataOut.data_cspc, |
|
|
1614 | 1614 | self.dataOut.data_dc) |
|
|
1615 | 1615 | else: |
|
|
1616 | 1616 | self.nProfiles=self.dataOut.nProfiles |
|
|
1617 | 1617 | self.nChannels=self.dataOut.nChannels |
|
|
1618 | 1618 | self.nHeights=self.dataOut.nHeights |
|
|
1619 | 1619 | avgdatatime, avgdata_spc, avgdata_cspc, avgdata_dc = self.integrate(self.dataOut.utctime, |
|
|
1620 | 1620 | self.dataOut.dataLag_spc, |
|
|
1621 | 1621 | self.dataOut.dataLag_cspc, |
|
|
1622 | 1622 | self.dataOut.dataLag_dc) |
|
|
1623 | 1623 | self.dataOut.flagNoData = True |
|
|
1624 | 1624 | if self.__dataReady: |
|
|
1625 | 1625 | |
|
|
1626 | 1626 | if not self.ByLags: |
|
|
1627 | 1627 | if self.nChannels == 1: |
|
|
1628 | 1628 | #print("f int", avgdata_spc.shape) |
|
|
1629 | 1629 | self.dataOut.data_spc = avgdata_spc |
|
|
1630 | 1630 | self.dataOut.data_cspc = avgdata_spc |
|
|
1631 | 1631 | else: |
|
|
1632 | 1632 | self.dataOut.data_spc = numpy.squeeze(avgdata_spc) |
|
|
1633 | 1633 | self.dataOut.data_cspc = numpy.squeeze(avgdata_cspc) |
|
|
1634 | 1634 | self.dataOut.data_dc = avgdata_dc |
|
|
1635 | 1635 | self.dataOut.data_outlier = self.dataOutliers |
|
|
1636 | 1636 | |
|
|
1637 | 1637 | else: |
|
|
1638 | 1638 | self.dataOut.dataLag_spc = avgdata_spc |
|
|
1639 | 1639 | self.dataOut.dataLag_cspc = avgdata_cspc |
|
|
1640 | 1640 | self.dataOut.dataLag_dc = avgdata_dc |
|
|
1641 | 1641 | |
|
|
1642 | 1642 | self.dataOut.data_spc=self.dataOut.dataLag_spc[:,:,:,self.dataOut.LagPlot] |
|
|
1643 | 1643 | self.dataOut.data_cspc=self.dataOut.dataLag_cspc[:,:,:,self.dataOut.LagPlot] |
|
|
1644 | 1644 | self.dataOut.data_dc=self.dataOut.dataLag_dc[:,:,self.dataOut.LagPlot] |
|
|
1645 | 1645 | |
|
|
1646 | 1646 | |
|
|
1647 | 1647 | self.dataOut.nIncohInt *= self.n_ints |
|
|
1648 | 1648 | self.dataOut.max_nIncohInt = self.maxProfilesInt |
|
|
1649 | 1649 | #print(self.dataOut.max_nIncohInt) |
|
|
1650 | 1650 | self.dataOut.utctime = avgdatatime |
|
|
1651 | 1651 | self.dataOut.flagNoData = False |
|
|
1652 | 1652 | #print("Faraday Integration DONE...") |
|
|
1653 | 1653 | #print(self.dataOut.flagNoData) |
|
|
1654 | 1654 | return self.dataOut |
|
|
1655 | 1655 | |
|
|
1656 | 1656 | class removeInterference(Operation): |
|
|
1657 | 1657 | |
|
|
1658 | 1658 | def removeInterference2(self): |
|
|
1659 | 1659 | |
|
|
1660 | 1660 | cspc = self.dataOut.data_cspc |
|
|
1661 | 1661 | spc = self.dataOut.data_spc |
|
|
1662 | 1662 | Heights = numpy.arange(cspc.shape[2]) |
|
|
1663 | 1663 | realCspc = numpy.abs(cspc) |
|
|
1664 | 1664 | |
|
|
1665 | 1665 | for i in range(cspc.shape[0]): |
|
|
1666 | 1666 | LinePower= numpy.sum(realCspc[i], axis=0) |
|
|
1667 | 1667 | Threshold = numpy.amax(LinePower)-numpy.sort(LinePower)[len(Heights)-int(len(Heights)*0.1)] |
|
|
1668 | 1668 | SelectedHeights = Heights[ numpy.where( LinePower < Threshold ) ] |
|
|
1669 | 1669 | InterferenceSum = numpy.sum( realCspc[i,:,SelectedHeights], axis=0 ) |
|
|
1670 | 1670 | InterferenceThresholdMin = numpy.sort(InterferenceSum)[int(len(InterferenceSum)*0.98)] |
|
|
1671 | 1671 | InterferenceThresholdMax = numpy.sort(InterferenceSum)[int(len(InterferenceSum)*0.99)] |
|
|
1672 | 1672 | |
|
|
1673 | 1673 | |
|
|
1674 | 1674 | InterferenceRange = numpy.where( ([InterferenceSum > InterferenceThresholdMin]))# , InterferenceSum < InterferenceThresholdMax]) ) |
|
|
1675 | 1675 | #InterferenceRange = numpy.where( ([InterferenceRange < InterferenceThresholdMax])) |
|
|
1676 | 1676 | if len(InterferenceRange)<int(cspc.shape[1]*0.3): |
|
|
1677 | 1677 | cspc[i,InterferenceRange,:] = numpy.NaN |
|
|
1678 | 1678 | |
|
|
1679 | 1679 | self.dataOut.data_cspc = cspc |
|
|
1680 | 1680 | |
|
|
1681 | 1681 | def removeInterference(self, interf = 2, hei_interf = None, nhei_interf = None, offhei_interf = None): |
|
|
1682 | 1682 | |
|
|
1683 | 1683 | jspectra = self.dataOut.data_spc |
|
|
1684 | 1684 | jcspectra = self.dataOut.data_cspc |
|
|
1685 | 1685 | jnoise = self.dataOut.getNoise() |
|
|
1686 | 1686 | num_incoh = self.dataOut.nIncohInt |
|
|
1687 | 1687 | |
|
|
1688 | 1688 | num_channel = jspectra.shape[0] |
|
|
1689 | 1689 | num_prof = jspectra.shape[1] |
|
|
1690 | 1690 | num_hei = jspectra.shape[2] |
|
|
1691 | 1691 | |
|
|
1692 | 1692 | # hei_interf |
|
|
1693 | 1693 | if hei_interf is None: |
|
|
1694 | 1694 | count_hei = int(num_hei / 2) |
|
|
1695 | 1695 | hei_interf = numpy.asmatrix(list(range(count_hei))) + num_hei - count_hei |
|
|
1696 | 1696 | hei_interf = numpy.asarray(hei_interf)[0] |
|
|
1697 | 1697 | # nhei_interf |
|
|
1698 | 1698 | if (nhei_interf == None): |
|
|
1699 | 1699 | nhei_interf = 5 |
|
|
1700 | 1700 | if (nhei_interf < 1): |
|
|
1701 | 1701 | nhei_interf = 1 |
|
|
1702 | 1702 | if (nhei_interf > count_hei): |
|
|
1703 | 1703 | nhei_interf = count_hei |
|
|
1704 | 1704 | if (offhei_interf == None): |
|
|
1705 | 1705 | offhei_interf = 0 |
|
|
1706 | 1706 | |
|
|
1707 | 1707 | ind_hei = list(range(num_hei)) |
|
|
1708 | 1708 | # mask_prof = numpy.asarray(range(num_prof - 2)) + 1 |
|
|
1709 | 1709 | # mask_prof[range(num_prof/2 - 1,len(mask_prof))] += 1 |
|
|
1710 | 1710 | mask_prof = numpy.asarray(list(range(num_prof))) |
|
|
1711 | 1711 | num_mask_prof = mask_prof.size |
|
|
1712 | 1712 | comp_mask_prof = [0, num_prof / 2] |
|
|
1713 | 1713 | |
|
|
1714 | 1714 | # noise_exist: Determina si la variable jnoise ha sido definida y contiene la informacion del ruido de cada canal |
|
|
1715 | 1715 | if (jnoise.size < num_channel or numpy.isnan(jnoise).any()): |
|
|
1716 | 1716 | jnoise = numpy.nan |
|
|
1717 | 1717 | noise_exist = jnoise[0] < numpy.Inf |
|
|
1718 | 1718 | |
|
|
1719 | 1719 | # Subrutina de Remocion de la Interferencia |
|
|
1720 | 1720 | for ich in range(num_channel): |
|
|
1721 | 1721 | # Se ordena los espectros segun su potencia (menor a mayor) |
|
|
1722 | 1722 | power = jspectra[ich, mask_prof, :] |
|
|
1723 | 1723 | power = power[:, hei_interf] |
|
|
1724 | 1724 | power = power.sum(axis=0) |
|
|
1725 | 1725 | psort = power.ravel().argsort() |
|
|
1726 | 1726 | |
|
|
1727 | 1727 | # Se estima la interferencia promedio en los Espectros de Potencia empleando |
|
|
1728 | 1728 | junkspc_interf = jspectra[ich, :, hei_interf[psort[list(range( |
|
|
1729 | 1729 | offhei_interf, nhei_interf + offhei_interf))]]] |
|
|
1730 | 1730 | |
|
|
1731 | 1731 | if noise_exist: |
|
|
1732 | 1732 | # tmp_noise = jnoise[ich] / num_prof |
|
|
1733 | 1733 | tmp_noise = jnoise[ich] |
|
|
1734 | 1734 | junkspc_interf = junkspc_interf - tmp_noise |
|
|
1735 | 1735 | #junkspc_interf[:,comp_mask_prof] = 0 |
|
|
1736 | 1736 | |
|
|
1737 | 1737 | jspc_interf = junkspc_interf.sum(axis=0) / nhei_interf |
|
|
1738 | 1738 | jspc_interf = jspc_interf.transpose() |
|
|
1739 | 1739 | # Calculando el espectro de interferencia promedio |
|
|
1740 | 1740 | noiseid = numpy.where( |
|
|
1741 | 1741 | jspc_interf <= tmp_noise / numpy.sqrt(num_incoh)) |
|
|
1742 | 1742 | noiseid = noiseid[0] |
|
|
1743 | 1743 | cnoiseid = noiseid.size |
|
|
1744 | 1744 | interfid = numpy.where( |
|
|
1745 | 1745 | jspc_interf > tmp_noise / numpy.sqrt(num_incoh)) |
|
|
1746 | 1746 | interfid = interfid[0] |
|
|
1747 | 1747 | cinterfid = interfid.size |
|
|
1748 | 1748 | |
|
|
1749 | 1749 | if (cnoiseid > 0): |
|
|
1750 | 1750 | jspc_interf[noiseid] = 0 |
|
|
1751 | 1751 | |
|
|
1752 | 1752 | # Expandiendo los perfiles a limpiar |
|
|
1753 | 1753 | if (cinterfid > 0): |
|
|
1754 | 1754 | new_interfid = ( |
|
|
1755 | 1755 | numpy.r_[interfid - 1, interfid, interfid + 1] + num_prof) % num_prof |
|
|
1756 | 1756 | new_interfid = numpy.asarray(new_interfid) |
|
|
1757 | 1757 | new_interfid = {x for x in new_interfid} |
|
|
1758 | 1758 | new_interfid = numpy.array(list(new_interfid)) |
|
|
1759 | 1759 | new_cinterfid = new_interfid.size |
|
|
1760 | 1760 | else: |
|
|
1761 | 1761 | new_cinterfid = 0 |
|
|
1762 | 1762 | |
|
|
1763 | 1763 | for ip in range(new_cinterfid): |
|
|
1764 | 1764 | ind = junkspc_interf[:, new_interfid[ip]].ravel().argsort() |
|
|
1765 | 1765 | jspc_interf[new_interfid[ip] |
|
|
1766 | 1766 | ] = junkspc_interf[ind[nhei_interf // 2], new_interfid[ip]] |
|
|
1767 | 1767 | |
|
|
1768 | 1768 | jspectra[ich, :, ind_hei] = jspectra[ich, :, |
|
|
1769 | 1769 | ind_hei] - jspc_interf # Corregir indices |
|
|
1770 | 1770 | |
|
|
1771 | 1771 | # Removiendo la interferencia del punto de mayor interferencia |
|
|
1772 | 1772 | ListAux = jspc_interf[mask_prof].tolist() |
|
|
1773 | 1773 | maxid = ListAux.index(max(ListAux)) |
|
|
1774 | 1774 | |
|
|
1775 | 1775 | if cinterfid > 0: |
|
|
1776 | 1776 | for ip in range(cinterfid * (interf == 2) - 1): |
|
|
1777 | 1777 | ind = (jspectra[ich, interfid[ip], :] < tmp_noise * |
|
|
1778 | 1778 | (1 + 1 / numpy.sqrt(num_incoh))).nonzero() |
|
|
1779 | 1779 | cind = len(ind) |
|
|
1780 | 1780 | |
|
|
1781 | 1781 | if (cind > 0): |
|
|
1782 | 1782 | jspectra[ich, interfid[ip], ind] = tmp_noise * \ |
|
|
1783 | 1783 | (1 + (numpy.random.uniform(cind) - 0.5) / |
|
|
1784 | 1784 | numpy.sqrt(num_incoh)) |
|
|
1785 | 1785 | |
|
|
1786 | 1786 | ind = numpy.array([-2, -1, 1, 2]) |
|
|
1787 | 1787 | xx = numpy.zeros([4, 4]) |
|
|
1788 | 1788 | |
|
|
1789 | 1789 | for id1 in range(4): |
|
|
1790 | 1790 | xx[:, id1] = ind[id1]**numpy.asarray(list(range(4))) |
|
|
1791 | 1791 | |
|
|
1792 | 1792 | xx_inv = numpy.linalg.inv(xx) |
|
|
1793 | 1793 | xx = xx_inv[:, 0] |
|
|
1794 | 1794 | ind = (ind + maxid + num_mask_prof) % num_mask_prof |
|
|
1795 | 1795 | yy = jspectra[ich, mask_prof[ind], :] |
|
|
1796 | 1796 | jspectra[ich, mask_prof[maxid], :] = numpy.dot( |
|
|
1797 | 1797 | yy.transpose(), xx) |
|
|
1798 | 1798 | |
|
|
1799 | 1799 | indAux = (jspectra[ich, :, :] < tmp_noise * |
|
|
1800 | 1800 | (1 - 1 / numpy.sqrt(num_incoh))).nonzero() |
|
|
1801 | 1801 | jspectra[ich, indAux[0], indAux[1]] = tmp_noise * \ |
|
|
1802 | 1802 | (1 - 1 / numpy.sqrt(num_incoh)) |
|
|
1803 | 1803 | |
|
|
1804 | 1804 | # Remocion de Interferencia en el Cross Spectra |
|
|
1805 | 1805 | if jcspectra is None: |
|
|
1806 | 1806 | return jspectra, jcspectra |
|
|
1807 | 1807 | num_pairs = int(jcspectra.size / (num_prof * num_hei)) |
|
|
1808 | 1808 | jcspectra = jcspectra.reshape(num_pairs, num_prof, num_hei) |
|
|
1809 | 1809 | |
|
|
1810 | 1810 | for ip in range(num_pairs): |
|
|
1811 | 1811 | |
|
|
1812 | 1812 | #------------------------------------------- |
|
|
1813 | 1813 | |
|
|
1814 | 1814 | cspower = numpy.abs(jcspectra[ip, mask_prof, :]) |
|
|
1815 | 1815 | cspower = cspower[:, hei_interf] |
|
|
1816 | 1816 | cspower = cspower.sum(axis=0) |
|
|
1817 | 1817 | |
|
|
1818 | 1818 | cspsort = cspower.ravel().argsort() |
|
|
1819 | 1819 | junkcspc_interf = jcspectra[ip, :, hei_interf[cspsort[list(range( |
|
|
1820 | 1820 | offhei_interf, nhei_interf + offhei_interf))]]] |
|
|
1821 | 1821 | junkcspc_interf = junkcspc_interf.transpose() |
|
|
1822 | 1822 | jcspc_interf = junkcspc_interf.sum(axis=1) / nhei_interf |
|
|
1823 | 1823 | |
|
|
1824 | 1824 | ind = numpy.abs(jcspc_interf[mask_prof]).ravel().argsort() |
|
|
1825 | 1825 | |
|
|
1826 | 1826 | median_real = int(numpy.median(numpy.real( |
|
|
1827 | 1827 | junkcspc_interf[mask_prof[ind[list(range(3 * num_prof // 4))]], :]))) |
|
|
1828 | 1828 | median_imag = int(numpy.median(numpy.imag( |
|
|
1829 | 1829 | junkcspc_interf[mask_prof[ind[list(range(3 * num_prof // 4))]], :]))) |
|
|
1830 | 1830 | comp_mask_prof = [int(e) for e in comp_mask_prof] |
|
|
1831 | 1831 | junkcspc_interf[comp_mask_prof, :] = numpy.complex( |
|
|
1832 | 1832 | median_real, median_imag) |
|
|
1833 | 1833 | |
|
|
1834 | 1834 | for iprof in range(num_prof): |
|
|
1835 | 1835 | ind = numpy.abs(junkcspc_interf[iprof, :]).ravel().argsort() |
|
|
1836 | 1836 | jcspc_interf[iprof] = junkcspc_interf[iprof, ind[nhei_interf // 2]] |
|
|
1837 | 1837 | |
|
|
1838 | 1838 | # Removiendo la Interferencia |
|
|
1839 | 1839 | jcspectra[ip, :, ind_hei] = jcspectra[ip, |
|
|
1840 | 1840 | :, ind_hei] - jcspc_interf |
|
|
1841 | 1841 | |
|
|
1842 | 1842 | ListAux = numpy.abs(jcspc_interf[mask_prof]).tolist() |
|
|
1843 | 1843 | maxid = ListAux.index(max(ListAux)) |
|
|
1844 | 1844 | |
|
|
1845 | 1845 | ind = numpy.array([-2, -1, 1, 2]) |
|
|
1846 | 1846 | xx = numpy.zeros([4, 4]) |
|
|
1847 | 1847 | |
|
|
1848 | 1848 | for id1 in range(4): |
|
|
1849 | 1849 | xx[:, id1] = ind[id1]**numpy.asarray(list(range(4))) |
|
|
1850 | 1850 | |
|
|
1851 | 1851 | xx_inv = numpy.linalg.inv(xx) |
|
|
1852 | 1852 | xx = xx_inv[:, 0] |
|
|
1853 | 1853 | |
|
|
1854 | 1854 | ind = (ind + maxid + num_mask_prof) % num_mask_prof |
|
|
1855 | 1855 | yy = jcspectra[ip, mask_prof[ind], :] |
|
|
1856 | 1856 | jcspectra[ip, mask_prof[maxid], :] = numpy.dot(yy.transpose(), xx) |
|
|
1857 | 1857 | |
|
|
1858 | 1858 | # Guardar Resultados |
|
|
1859 | 1859 | self.dataOut.data_spc = jspectra |
|
|
1860 | 1860 | self.dataOut.data_cspc = jcspectra |
|
|
1861 | 1861 | |
|
|
1862 | 1862 | return 1 |
|
|
1863 | 1863 | |
|
|
1864 | 1864 | def run(self, dataOut, interf = 2,hei_interf = None, nhei_interf = None, offhei_interf = None, mode=1): |
|
|
1865 | 1865 | |
|
|
1866 | 1866 | self.dataOut = dataOut |
|
|
1867 | 1867 | |
|
|
1868 | 1868 | if mode == 1: |
|
|
1869 | 1869 | self.removeInterference(interf = 2,hei_interf = None, nhei_interf = None, offhei_interf = None) |
|
|
1870 | 1870 | elif mode == 2: |
|
|
1871 | 1871 | self.removeInterference2() |
|
|
1872 | 1872 | |
|
|
1873 | 1873 | return self.dataOut |
|
|
1874 | 1874 | |
|
|
1875 | 1875 | |
|
|
1876 | 1876 | class IncohInt(Operation): |
|
|
1877 | 1877 | |
|
|
1878 | 1878 | __profIndex = 0 |
|
|
1879 | 1879 | __withOverapping = False |
|
|
1880 | 1880 | |
|
|
1881 | 1881 | __byTime = False |
|
|
1882 | 1882 | __initime = None |
|
|
1883 | 1883 | __lastdatatime = None |
|
|
1884 | 1884 | __integrationtime = None |
|
|
1885 | 1885 | |
|
|
1886 | 1886 | __buffer_spc = None |
|
|
1887 | 1887 | __buffer_cspc = None |
|
|
1888 | 1888 | __buffer_dc = None |
|
|
1889 | 1889 | |
|
|
1890 | 1890 | __dataReady = False |
|
|
1891 | 1891 | |
|
|
1892 | 1892 | __timeInterval = None |
|
|
1893 | 1893 | incohInt = 0 |
|
|
1894 | 1894 | nOutliers = 0 |
|
|
1895 | 1895 | n = None |
|
|
1896 | 1896 | |
|
|
1897 | 1897 | def __init__(self): |
|
|
1898 | 1898 | |
|
|
1899 | 1899 | Operation.__init__(self) |
|
|
1900 | 1900 | |
|
|
1901 | 1901 | def setup(self, n=None, timeInterval=None, overlapping=False): |
|
|
1902 | 1902 | """ |
|
|
1903 | 1903 | Set the parameters of the integration class. |
|
|
1904 | 1904 | |
|
|
1905 | 1905 | Inputs: |
|
|
1906 | 1906 | |
|
|
1907 | 1907 | n : Number of coherent integrations |
|
|
1908 | 1908 | timeInterval : Time of integration. If the parameter "n" is selected this one does not work |
|
|
1909 | 1909 | overlapping : |
|
|
1910 | 1910 | |
|
|
1911 | 1911 | """ |
|
|
1912 | 1912 | |
|
|
1913 | 1913 | self.__initime = None |
|
|
1914 | 1914 | self.__lastdatatime = 0 |
|
|
1915 | 1915 | |
|
|
1916 | 1916 | self.__buffer_spc = 0 |
|
|
1917 | 1917 | self.__buffer_cspc = 0 |
|
|
1918 | 1918 | self.__buffer_dc = 0 |
|
|
1919 | 1919 | |
|
|
1920 | 1920 | self.__profIndex = 0 |
|
|
1921 | 1921 | self.__dataReady = False |
|
|
1922 | 1922 | self.__byTime = False |
|
|
1923 | 1923 | self.incohInt = 0 |
|
|
1924 | 1924 | self.nOutliers = 0 |
|
|
1925 | 1925 | if n is None and timeInterval is None: |
|
|
1926 | 1926 | raise ValueError("n or timeInterval should be specified ...") |
|
|
1927 | 1927 | |
|
|
1928 | 1928 | if n is not None: |
|
|
1929 | 1929 | self.n = int(n) |
|
|
1930 | 1930 | else: |
|
|
1931 | 1931 | |
|
|
1932 | 1932 | self.__integrationtime = int(timeInterval) |
|
|
1933 | 1933 | self.n = None |
|
|
1934 | 1934 | self.__byTime = True |
|
|
1935 | 1935 | |
|
|
1936 | 1936 | def putData(self, data_spc, data_cspc, data_dc): |
|
|
1937 | 1937 | """ |
|
|
1938 | 1938 | Add a profile to the __buffer_spc and increase in one the __profileIndex |
|
|
1939 | 1939 | |
|
|
1940 | 1940 | """ |
|
|
1941 | 1941 | if data_spc.all() == numpy.nan : |
|
|
1942 | 1942 | print("nan ") |
|
|
1943 | 1943 | return |
|
|
1944 | 1944 | self.__buffer_spc += data_spc |
|
|
1945 | 1945 | |
|
|
1946 | 1946 | if data_cspc is None: |
|
|
1947 | 1947 | self.__buffer_cspc = None |
|
|
1948 | 1948 | else: |
|
|
1949 | 1949 | self.__buffer_cspc += data_cspc |
|
|
1950 | 1950 | |
|
|
1951 | 1951 | if data_dc is None: |
|
|
1952 | 1952 | self.__buffer_dc = None |
|
|
1953 | 1953 | else: |
|
|
1954 | 1954 | self.__buffer_dc += data_dc |
|
|
1955 | 1955 | |
|
|
1956 | 1956 | self.__profIndex += 1 |
|
|
1957 | 1957 | |
|
|
1958 | 1958 | return |
|
|
1959 | 1959 | |
|
|
1960 | 1960 | def pushData(self): |
|
|
1961 | 1961 | """ |
|
|
1962 | 1962 | Return the sum of the last profiles and the profiles used in the sum. |
|
|
1963 | 1963 | |
|
|
1964 | 1964 | Affected: |
|
|
1965 | 1965 | |
|
|
1966 | 1966 | self.__profileIndex |
|
|
1967 | 1967 | |
|
|
1968 | 1968 | """ |
|
|
1969 | 1969 | |
|
|
1970 | 1970 | data_spc = self.__buffer_spc |
|
|
1971 | 1971 | data_cspc = self.__buffer_cspc |
|
|
1972 | 1972 | data_dc = self.__buffer_dc |
|
|
1973 | 1973 | n = self.__profIndex |
|
|
1974 | 1974 | |
|
|
1975 | 1975 | self.__buffer_spc = 0 |
|
|
1976 | 1976 | self.__buffer_cspc = 0 |
|
|
1977 | 1977 | self.__buffer_dc = 0 |
|
|
1978 | 1978 | |
|
|
1979 | 1979 | |
|
|
1980 | 1980 | return data_spc, data_cspc, data_dc, n |
|
|
1981 | 1981 | |
|
|
1982 | 1982 | def byProfiles(self, *args): |
|
|
1983 | 1983 | |
|
|
1984 | 1984 | self.__dataReady = False |
|
|
1985 | 1985 | avgdata_spc = None |
|
|
1986 | 1986 | avgdata_cspc = None |
|
|
1987 | 1987 | avgdata_dc = None |
|
|
1988 | 1988 | |
|
|
1989 | 1989 | self.putData(*args) |
|
|
1990 | 1990 | |
|
|
1991 | 1991 | if self.__profIndex == self.n: |
|
|
1992 | 1992 | |
|
|
1993 | 1993 | avgdata_spc, avgdata_cspc, avgdata_dc, n = self.pushData() |
|
|
1994 | 1994 | self.n = n |
|
|
1995 | 1995 | self.__dataReady = True |
|
|
1996 | 1996 | |
|
|
1997 | 1997 | return avgdata_spc, avgdata_cspc, avgdata_dc |
|
|
1998 | 1998 | |
|
|
1999 | 1999 | def byTime(self, datatime, *args): |
|
|
2000 | 2000 | |
|
|
2001 | 2001 | self.__dataReady = False |
|
|
2002 | 2002 | avgdata_spc = None |
|
|
2003 | 2003 | avgdata_cspc = None |
|
|
2004 | 2004 | avgdata_dc = None |
|
|
2005 | 2005 | |
|
|
2006 | 2006 | self.putData(*args) |
|
|
2007 | 2007 | |
|
|
2008 | 2008 | if (datatime - self.__initime) >= self.__integrationtime: |
|
|
2009 | 2009 | avgdata_spc, avgdata_cspc, avgdata_dc, n = self.pushData() |
|
|
2010 | 2010 | self.n = n |
|
|
2011 | 2011 | self.__dataReady = True |
|
|
2012 | 2012 | |
|
|
2013 | 2013 | return avgdata_spc, avgdata_cspc, avgdata_dc |
|
|
2014 | 2014 | |
|
|
2015 | 2015 | def integrate(self, datatime, *args): |
|
|
2016 | 2016 | |
|
|
2017 | 2017 | if self.__profIndex == 0: |
|
|
2018 | 2018 | self.__initime = datatime |
|
|
2019 | 2019 | |
|
|
2020 | 2020 | if self.__byTime: |
|
|
2021 | 2021 | avgdata_spc, avgdata_cspc, avgdata_dc = self.byTime( |
|
|
2022 | 2022 | datatime, *args) |
|
|
2023 | 2023 | else: |
|
|
2024 | 2024 | avgdata_spc, avgdata_cspc, avgdata_dc = self.byProfiles(*args) |
|
|
2025 | 2025 | |
|
|
2026 | 2026 | if not self.__dataReady: |
|
|
2027 | 2027 | return None, None, None, None |
|
|
2028 | 2028 | |
|
|
2029 | 2029 | return self.__initime, avgdata_spc, avgdata_cspc, avgdata_dc |
|
|
2030 | 2030 | |
|
|
2031 | 2031 | def run(self, dataOut, n=None, timeInterval=None, overlapping=False): |
|
|
2032 | 2032 | if n == 1: |
|
|
2033 | 2033 | return dataOut |
|
|
2034 | 2034 | |
|
|
2035 | 2035 | if dataOut.flagNoData == True: |
|
|
2036 | 2036 | return dataOut |
|
|
2037 | 2037 | |
|
|
2038 | 2038 | dataOut.flagNoData = True |
|
|
2039 | 2039 | |
|
|
2040 | 2040 | if not self.isConfig: |
|
|
2041 | 2041 | self.setup(n, timeInterval, overlapping) |
|
|
2042 | 2042 | self.isConfig = True |
|
|
2043 | 2043 | |
|
|
2044 | 2044 | avgdatatime, avgdata_spc, avgdata_cspc, avgdata_dc = self.integrate(dataOut.utctime, |
|
|
2045 | 2045 | dataOut.data_spc, |
|
|
2046 | 2046 | dataOut.data_cspc, |
|
|
2047 | 2047 | dataOut.data_dc) |
|
|
2048 | 2048 | self.incohInt += dataOut.nIncohInt |
|
|
2049 | 2049 | self.nOutliers += dataOut.data_outlier |
|
|
2050 | 2050 | if self.__dataReady: |
|
|
2051 | 2051 | #print("prof: ",dataOut.max_nIncohInt,self.__profIndex) |
|
|
2052 | 2052 | dataOut.data_spc = avgdata_spc |
|
|
2053 | 2053 | dataOut.data_cspc = avgdata_cspc |
|
|
2054 | 2054 | dataOut.data_dc = avgdata_dc |
|
|
2055 | 2055 | dataOut.nIncohInt = self.incohInt |
|
|
2056 | 2056 | dataOut.data_outlier = self.nOutliers |
|
|
2057 | 2057 | dataOut.utctime = avgdatatime |
|
|
2058 | 2058 | dataOut.flagNoData = False |
|
|
2059 | 2059 | dataOut.max_nIncohInt += self.__profIndex |
|
|
2060 | 2060 | self.incohInt = 0 |
|
|
2061 | 2061 | self.nOutliers = 0 |
|
|
2062 | 2062 | self.__profIndex = 0 |
|
|
2063 | 2063 | #print("IncohInt Done") |
|
|
2064 | 2064 | return dataOut |
|
|
2065 | 2065 | |
|
|
2066 | 2066 | class dopplerFlip(Operation): |
|
|
2067 | 2067 | |
|
|
2068 | 2068 | def run(self, dataOut): |
|
|
2069 | 2069 | # arreglo 1: (num_chan, num_profiles, num_heights) |
|
|
2070 | 2070 | self.dataOut = dataOut |
|
|
2071 | 2071 | # JULIA-oblicua, indice 2 |
|
|
2072 | 2072 | # arreglo 2: (num_profiles, num_heights) |
|
|
2073 | 2073 | jspectra = self.dataOut.data_spc[2] |
|
|
2074 | 2074 | jspectra_tmp = numpy.zeros(jspectra.shape) |
|
|
2075 | 2075 | num_profiles = jspectra.shape[0] |
|
|
2076 | 2076 | freq_dc = int(num_profiles / 2) |
|
|
2077 | 2077 | # Flip con for |
|
|
2078 | 2078 | for j in range(num_profiles): |
|
|
2079 | 2079 | jspectra_tmp[num_profiles-j-1]= jspectra[j] |
|
|
2080 | 2080 | # Intercambio perfil de DC con perfil inmediato anterior |
|
|
2081 | 2081 | jspectra_tmp[freq_dc-1]= jspectra[freq_dc-1] |
|
|
2082 | 2082 | jspectra_tmp[freq_dc]= jspectra[freq_dc] |
|
|
2083 | 2083 | # canal modificado es re-escrito en el arreglo de canales |
|
|
2084 | 2084 | self.dataOut.data_spc[2] = jspectra_tmp |
|
|
2085 | 2085 | |
|
|
2086 | 2086 | return self.dataOut |
| @@ -1,2361 +1,2367 | |||
|
|
1 | 1 | import sys |
|
|
2 | 2 | import numpy,math |
|
|
3 | 3 | from scipy import interpolate |
|
|
4 | 4 | from schainpy.model.proc.jroproc_base import ProcessingUnit, Operation, MPDecorator |
|
|
5 | 5 | from schainpy.model.data.jrodata import Voltage,hildebrand_sekhon |
|
|
6 | 6 | from schainpy.utils import log |
|
|
7 | 7 | from schainpy.model.io.utils import getHei_index |
|
|
8 | 8 | from time import time |
|
|
9 | #import datetime | |
|
|
9 | ||
|
|
10 | 10 | import numpy |
|
|
11 | 11 | #import copy |
|
|
12 | 12 | from schainpy.model.data import _noise |
|
|
13 | 13 | |
|
|
14 | 14 | class VoltageProc(ProcessingUnit): |
|
|
15 | 15 | |
|
|
16 | 16 | def __init__(self): |
|
|
17 | 17 | |
|
|
18 | 18 | ProcessingUnit.__init__(self) |
|
|
19 | 19 | |
|
|
20 | 20 | self.dataOut = Voltage() |
|
|
21 | 21 | self.flip = 1 |
|
|
22 | 22 | self.setupReq = False |
|
|
23 | 23 | |
|
|
24 | 24 | def run(self): |
|
|
25 | 25 | #print("running volt proc") |
|
|
26 | 26 | if self.dataIn.type == 'AMISR': |
|
|
27 | 27 | self.__updateObjFromAmisrInput() |
|
|
28 | 28 | |
|
|
29 | 29 | if self.dataOut.buffer_empty: |
|
|
30 | 30 | if self.dataIn.type == 'Voltage': |
|
|
31 | 31 | self.dataOut.copy(self.dataIn) |
|
|
32 | 32 | #print("new volts reading") |
|
|
33 | 33 | |
|
|
34 | 34 | |
|
|
35 | 35 | def __updateObjFromAmisrInput(self): |
|
|
36 | 36 | |
|
|
37 | 37 | self.dataOut.timeZone = self.dataIn.timeZone |
|
|
38 | 38 | self.dataOut.dstFlag = self.dataIn.dstFlag |
|
|
39 | 39 | self.dataOut.errorCount = self.dataIn.errorCount |
|
|
40 | 40 | self.dataOut.useLocalTime = self.dataIn.useLocalTime |
|
|
41 | 41 | |
|
|
42 | 42 | self.dataOut.flagNoData = self.dataIn.flagNoData |
|
|
43 | 43 | self.dataOut.data = self.dataIn.data |
|
|
44 | 44 | self.dataOut.utctime = self.dataIn.utctime |
|
|
45 | 45 | self.dataOut.channelList = self.dataIn.channelList |
|
|
46 | 46 | #self.dataOut.timeInterval = self.dataIn.timeInterval |
|
|
47 | 47 | self.dataOut.heightList = self.dataIn.heightList |
|
|
48 | 48 | self.dataOut.nProfiles = self.dataIn.nProfiles |
|
|
49 | 49 | |
|
|
50 | 50 | self.dataOut.nCohInt = self.dataIn.nCohInt |
|
|
51 | 51 | self.dataOut.ippSeconds = self.dataIn.ippSeconds |
|
|
52 | 52 | self.dataOut.frequency = self.dataIn.frequency |
|
|
53 | 53 | |
|
|
54 | 54 | self.dataOut.azimuth = self.dataIn.azimuth |
|
|
55 | 55 | self.dataOut.zenith = self.dataIn.zenith |
|
|
56 | 56 | |
|
|
57 | 57 | self.dataOut.beam.codeList = self.dataIn.beam.codeList |
|
|
58 | 58 | self.dataOut.beam.azimuthList = self.dataIn.beam.azimuthList |
|
|
59 | 59 | self.dataOut.beam.zenithList = self.dataIn.beam.zenithList |
|
|
60 | 60 | |
|
|
61 | 61 | |
|
|
62 | 62 | class selectChannels(Operation): |
|
|
63 | 63 | |
|
|
64 | 64 | def run(self, dataOut, channelList=None): |
|
|
65 | 65 | self.channelList = channelList |
|
|
66 | 66 | if self.channelList == None: |
|
|
67 | 67 | print("Missing channelList") |
|
|
68 | 68 | return dataOut |
|
|
69 | 69 | channelIndexList = [] |
|
|
70 | 70 | |
|
|
71 | 71 | if type(dataOut.channelList) is not list: #leer array desde HDF5 |
|
|
72 | 72 | try: |
|
|
73 | 73 | dataOut.channelList = dataOut.channelList.tolist() |
|
|
74 | 74 | except Exception as e: |
|
|
75 | 75 | print("Select Channels: ",e) |
|
|
76 | 76 | for channel in self.channelList: |
|
|
77 | 77 | if channel not in dataOut.channelList: |
|
|
78 | 78 | raise ValueError("Channel %d is not in %s" %(channel, str(dataOut.channelList))) |
|
|
79 | 79 | |
|
|
80 | 80 | index = dataOut.channelList.index(channel) |
|
|
81 | 81 | channelIndexList.append(index) |
|
|
82 | 82 | dataOut = self.selectChannelsByIndex(dataOut,channelIndexList) |
|
|
83 | 83 | return dataOut |
|
|
84 | 84 | |
|
|
85 | 85 | def selectChannelsByIndex(self, dataOut, channelIndexList): |
|
|
86 | 86 | """ |
|
|
87 | 87 | Selecciona un bloque de datos en base a canales segun el channelIndexList |
|
|
88 | 88 | |
|
|
89 | 89 | Input: |
|
|
90 | 90 | channelIndexList : lista sencilla de canales a seleccionar por ej. [2,3,7] |
|
|
91 | 91 | |
|
|
92 | 92 | Affected: |
|
|
93 | 93 | dataOut.data |
|
|
94 | 94 | dataOut.channelIndexList |
|
|
95 | 95 | dataOut.nChannels |
|
|
96 | 96 | dataOut.m_ProcessingHeader.totalSpectra |
|
|
97 | 97 | dataOut.systemHeaderObj.numChannels |
|
|
98 | 98 | dataOut.m_ProcessingHeader.blockSize |
|
|
99 | 99 | |
|
|
100 | 100 | Return: |
|
|
101 | 101 | None |
|
|
102 | 102 | """ |
|
|
103 | 103 | #print("selectChannelsByIndex") |
|
|
104 | 104 | # for channelIndex in channelIndexList: |
|
|
105 | 105 | # if channelIndex not in dataOut.channelIndexList: |
|
|
106 | 106 | # raise ValueError("The value %d in channelIndexList is not valid" %channelIndex) |
|
|
107 | 107 | |
|
|
108 | 108 | if dataOut.type == 'Voltage': |
|
|
109 | 109 | if dataOut.flagDataAsBlock: |
|
|
110 | 110 | """ |
|
|
111 | 111 | Si la data es obtenida por bloques, dimension = [nChannels, nProfiles, nHeis] |
|
|
112 | 112 | """ |
|
|
113 | 113 | data = dataOut.data[channelIndexList,:,:] |
|
|
114 | 114 | else: |
|
|
115 | 115 | data = dataOut.data[channelIndexList,:] |
|
|
116 | 116 | |
|
|
117 | 117 | dataOut.data = data |
|
|
118 | 118 | # dataOut.channelList = [dataOut.channelList[i] for i in channelIndexList] |
|
|
119 | 119 | dataOut.channelList = range(len(channelIndexList)) |
|
|
120 | 120 | |
|
|
121 | 121 | elif dataOut.type == 'Spectra': |
|
|
122 | 122 | if hasattr(dataOut, 'data_spc'): |
|
|
123 | 123 | if dataOut.data_spc is None: |
|
|
124 | 124 | raise ValueError("data_spc is None") |
|
|
125 | 125 | return dataOut |
|
|
126 | 126 | else: |
|
|
127 | 127 | data_spc = dataOut.data_spc[channelIndexList, :] |
|
|
128 | 128 | dataOut.data_spc = data_spc |
|
|
129 | 129 | |
|
|
130 | 130 | # if hasattr(dataOut, 'data_dc') :# and |
|
|
131 | 131 | # if dataOut.data_dc is None: |
|
|
132 | 132 | # raise ValueError("data_dc is None") |
|
|
133 | 133 | # return dataOut |
|
|
134 | 134 | # else: |
|
|
135 | 135 | # data_dc = dataOut.data_dc[channelIndexList, :] |
|
|
136 | 136 | # dataOut.data_dc = data_dc |
|
|
137 | 137 | # dataOut.channelList = [dataOut.channelList[i] for i in channelIndexList] |
|
|
138 | 138 | dataOut.channelList = channelIndexList |
|
|
139 | 139 | dataOut = self.__selectPairsByChannel(dataOut,channelIndexList) |
|
|
140 | 140 | |
|
|
141 | 141 | return dataOut |
|
|
142 | 142 | |
|
|
143 | 143 | def __selectPairsByChannel(self, dataOut, channelList=None): |
|
|
144 | 144 | #print("__selectPairsByChannel") |
|
|
145 | 145 | if channelList == None: |
|
|
146 | 146 | return |
|
|
147 | 147 | |
|
|
148 | 148 | pairsIndexListSelected = [] |
|
|
149 | 149 | for pairIndex in dataOut.pairsIndexList: |
|
|
150 | 150 | # First pair |
|
|
151 | 151 | if dataOut.pairsList[pairIndex][0] not in channelList: |
|
|
152 | 152 | continue |
|
|
153 | 153 | # Second pair |
|
|
154 | 154 | if dataOut.pairsList[pairIndex][1] not in channelList: |
|
|
155 | 155 | continue |
|
|
156 | 156 | |
|
|
157 | 157 | pairsIndexListSelected.append(pairIndex) |
|
|
158 | 158 | if not pairsIndexListSelected: |
|
|
159 | 159 | dataOut.data_cspc = None |
|
|
160 | 160 | dataOut.pairsList = [] |
|
|
161 | 161 | return |
|
|
162 | 162 | |
|
|
163 | 163 | dataOut.data_cspc = dataOut.data_cspc[pairsIndexListSelected] |
|
|
164 | 164 | dataOut.pairsList = [dataOut.pairsList[i] |
|
|
165 | 165 | for i in pairsIndexListSelected] |
|
|
166 | 166 | |
|
|
167 | 167 | return dataOut |
|
|
168 | 168 | |
|
|
169 | 169 | class selectHeights(Operation): |
|
|
170 | 170 | |
|
|
171 | 171 | def run(self, dataOut, minHei=None, maxHei=None, minIndex=None, maxIndex=None): |
|
|
172 | 172 | """ |
|
|
173 | 173 | Selecciona un bloque de datos en base a un grupo de valores de alturas segun el rango |
|
|
174 | 174 | minHei <= height <= maxHei |
|
|
175 | 175 | |
|
|
176 | 176 | Input: |
|
|
177 | 177 | minHei : valor minimo de altura a considerar |
|
|
178 | 178 | maxHei : valor maximo de altura a considerar |
|
|
179 | 179 | |
|
|
180 | 180 | Affected: |
|
|
181 | 181 | Indirectamente son cambiados varios valores a travez del metodo selectHeightsByIndex |
|
|
182 | 182 | |
|
|
183 | 183 | Return: |
|
|
184 | 184 | 1 si el metodo se ejecuto con exito caso contrario devuelve 0 |
|
|
185 | 185 | """ |
|
|
186 | 186 | |
|
|
187 | 187 | self.dataOut = dataOut |
|
|
188 | 188 | |
|
|
189 | 189 | if minHei and maxHei: |
|
|
190 | 190 | |
|
|
191 | 191 | if (minHei < dataOut.heightList[0]): |
|
|
192 | 192 | minHei = dataOut.heightList[0] |
|
|
193 | 193 | |
|
|
194 | 194 | if (maxHei > dataOut.heightList[-1]): |
|
|
195 | 195 | maxHei = dataOut.heightList[-1] |
|
|
196 | 196 | |
|
|
197 | 197 | minIndex = 0 |
|
|
198 | 198 | maxIndex = 0 |
|
|
199 | 199 | heights = dataOut.heightList |
|
|
200 | 200 | |
|
|
201 | 201 | inda = numpy.where(heights >= minHei) |
|
|
202 | 202 | indb = numpy.where(heights <= maxHei) |
|
|
203 | 203 | |
|
|
204 | 204 | try: |
|
|
205 | 205 | minIndex = inda[0][0] |
|
|
206 | 206 | except: |
|
|
207 | 207 | minIndex = 0 |
|
|
208 | 208 | |
|
|
209 | 209 | try: |
|
|
210 | 210 | maxIndex = indb[0][-1] |
|
|
211 | 211 | except: |
|
|
212 | 212 | maxIndex = len(heights) |
|
|
213 | 213 | |
|
|
214 | 214 | self.selectHeightsByIndex(minIndex, maxIndex) |
|
|
215 | 215 | |
|
|
216 | 216 | return dataOut |
|
|
217 | 217 | |
|
|
218 | 218 | def selectHeightsByIndex(self, minIndex, maxIndex): |
|
|
219 | 219 | """ |
|
|
220 | 220 | Selecciona un bloque de datos en base a un grupo indices de alturas segun el rango |
|
|
221 | 221 | minIndex <= index <= maxIndex |
|
|
222 | 222 | |
|
|
223 | 223 | Input: |
|
|
224 | 224 | minIndex : valor de indice minimo de altura a considerar |
|
|
225 | 225 | maxIndex : valor de indice maximo de altura a considerar |
|
|
226 | 226 | |
|
|
227 | 227 | Affected: |
|
|
228 | 228 | self.dataOut.data |
|
|
229 | 229 | self.dataOut.heightList |
|
|
230 | 230 | |
|
|
231 | 231 | Return: |
|
|
232 | 232 | 1 si el metodo se ejecuto con exito caso contrario devuelve 0 |
|
|
233 | 233 | """ |
|
|
234 | 234 | |
|
|
235 | 235 | if self.dataOut.type == 'Voltage': |
|
|
236 | 236 | if (minIndex < 0) or (minIndex > maxIndex): |
|
|
237 | 237 | raise ValueError("Height index range (%d,%d) is not valid" % (minIndex, maxIndex)) |
|
|
238 | 238 | |
|
|
239 | 239 | if (maxIndex >= self.dataOut.nHeights): |
|
|
240 | 240 | maxIndex = self.dataOut.nHeights |
|
|
241 | 241 | |
|
|
242 | 242 | #voltage |
|
|
243 | 243 | if self.dataOut.flagDataAsBlock: |
|
|
244 | 244 | """ |
|
|
245 | 245 | Si la data es obtenida por bloques, dimension = [nChannels, nProfiles, nHeis] |
|
|
246 | 246 | """ |
|
|
247 | 247 | data = self.dataOut.data[:,:, minIndex:maxIndex] |
|
|
248 | 248 | else: |
|
|
249 | 249 | data = self.dataOut.data[:, minIndex:maxIndex] |
|
|
250 | 250 | |
|
|
251 | 251 | # firstHeight = self.dataOut.heightList[minIndex] |
|
|
252 | 252 | |
|
|
253 | 253 | self.dataOut.data = data |
|
|
254 | 254 | self.dataOut.heightList = self.dataOut.heightList[minIndex:maxIndex] |
|
|
255 | 255 | |
|
|
256 | 256 | if self.dataOut.nHeights <= 1: |
|
|
257 | 257 | raise ValueError("selectHeights: Too few heights. Current number of heights is %d" %(self.dataOut.nHeights)) |
|
|
258 | 258 | elif self.dataOut.type == 'Spectra': |
|
|
259 | 259 | if (minIndex < 0) or (minIndex > maxIndex): |
|
|
260 | 260 | raise ValueError("Error selecting heights: Index range (%d,%d) is not valid" % ( |
|
|
261 | 261 | minIndex, maxIndex)) |
|
|
262 | 262 | |
|
|
263 | 263 | if (maxIndex >= self.dataOut.nHeights): |
|
|
264 | 264 | maxIndex = self.dataOut.nHeights - 1 |
|
|
265 | 265 | |
|
|
266 | 266 | # Spectra |
|
|
267 | 267 | data_spc = self.dataOut.data_spc[:, :, minIndex:maxIndex + 1] |
|
|
268 | 268 | |
|
|
269 | 269 | data_cspc = None |
|
|
270 | 270 | if self.dataOut.data_cspc is not None: |
|
|
271 | 271 | data_cspc = self.dataOut.data_cspc[:, :, minIndex:maxIndex + 1] |
|
|
272 | 272 | |
|
|
273 | 273 | data_dc = None |
|
|
274 | 274 | if self.dataOut.data_dc is not None: |
|
|
275 | 275 | data_dc = self.dataOut.data_dc[:, minIndex:maxIndex + 1] |
|
|
276 | 276 | |
|
|
277 | 277 | self.dataOut.data_spc = data_spc |
|
|
278 | 278 | self.dataOut.data_cspc = data_cspc |
|
|
279 | 279 | self.dataOut.data_dc = data_dc |
|
|
280 | 280 | |
|
|
281 | 281 | self.dataOut.heightList = self.dataOut.heightList[minIndex:maxIndex + 1] |
|
|
282 | 282 | |
|
|
283 | 283 | return 1 |
|
|
284 | 284 | |
|
|
285 | 285 | |
|
|
286 | 286 | class filterByHeights(Operation): |
|
|
287 | 287 | |
|
|
288 | 288 | def run(self, dataOut, window): |
|
|
289 | 289 | |
|
|
290 | 290 | deltaHeight = dataOut.heightList[1] - dataOut.heightList[0] |
|
|
291 | 291 | |
|
|
292 | 292 | if window == None: |
|
|
293 | 293 | window = (dataOut.radarControllerHeaderObj.txA/dataOut.radarControllerHeaderObj.nBaud) / deltaHeight |
|
|
294 | 294 | |
|
|
295 | 295 | newdelta = deltaHeight * window |
|
|
296 | 296 | r = dataOut.nHeights % window |
|
|
297 | 297 | newheights = (dataOut.nHeights-r)/window |
|
|
298 | 298 | |
|
|
299 | 299 | if newheights <= 1: |
|
|
300 | 300 | raise ValueError("filterByHeights: Too few heights. Current number of heights is %d and window is %d" %(dataOut.nHeights, window)) |
|
|
301 | 301 | |
|
|
302 | 302 | if dataOut.flagDataAsBlock: |
|
|
303 | 303 | """ |
|
|
304 | 304 | Si la data es obtenida por bloques, dimension = [nChannels, nProfiles, nHeis] |
|
|
305 | 305 | """ |
|
|
306 | 306 | buffer = dataOut.data[:, :, 0:int(dataOut.nHeights-r)] |
|
|
307 | 307 | buffer = buffer.reshape(dataOut.nChannels, dataOut.nProfiles, int(dataOut.nHeights/window), window) |
|
|
308 | 308 | buffer = numpy.sum(buffer,3) |
|
|
309 | 309 | |
|
|
310 | 310 | else: |
|
|
311 | 311 | buffer = dataOut.data[:,0:int(dataOut.nHeights-r)] |
|
|
312 | 312 | buffer = buffer.reshape(dataOut.nChannels,int(dataOut.nHeights/window),int(window)) |
|
|
313 | 313 | buffer = numpy.sum(buffer,2) |
|
|
314 | 314 | |
|
|
315 | 315 | dataOut.data = buffer |
|
|
316 | 316 | dataOut.heightList = dataOut.heightList[0] + numpy.arange( newheights )*newdelta |
|
|
317 | 317 | dataOut.windowOfFilter = window |
|
|
318 | 318 | |
|
|
319 | 319 | return dataOut |
|
|
320 | 320 | |
|
|
321 | 321 | |
|
|
322 | 322 | class setH0(Operation): |
|
|
323 | 323 | |
|
|
324 | 324 | def run(self, dataOut, h0, deltaHeight = None): |
|
|
325 | 325 | |
|
|
326 | 326 | if not deltaHeight: |
|
|
327 | 327 | deltaHeight = dataOut.heightList[1] - dataOut.heightList[0] |
|
|
328 | 328 | |
|
|
329 | 329 | nHeights = dataOut.nHeights |
|
|
330 | 330 | |
|
|
331 | 331 | newHeiRange = h0 + numpy.arange(nHeights)*deltaHeight |
|
|
332 | 332 | |
|
|
333 | 333 | dataOut.heightList = newHeiRange |
|
|
334 | 334 | |
|
|
335 | 335 | return dataOut |
|
|
336 | 336 | |
|
|
337 | 337 | |
|
|
338 | 338 | class deFlip(Operation): |
|
|
339 | 339 | |
|
|
340 | 340 | def run(self, dataOut, channelList = []): |
|
|
341 | 341 | |
|
|
342 | 342 | data = dataOut.data.copy() |
|
|
343 | 343 | |
|
|
344 | 344 | if dataOut.flagDataAsBlock: |
|
|
345 | 345 | flip = self.flip |
|
|
346 | 346 | profileList = list(range(dataOut.nProfiles)) |
|
|
347 | 347 | |
|
|
348 | 348 | if not channelList: |
|
|
349 | 349 | for thisProfile in profileList: |
|
|
350 | 350 | data[:,thisProfile,:] = data[:,thisProfile,:]*flip |
|
|
351 | 351 | flip *= -1.0 |
|
|
352 | 352 | else: |
|
|
353 | 353 | for thisChannel in channelList: |
|
|
354 | 354 | if thisChannel not in dataOut.channelList: |
|
|
355 | 355 | continue |
|
|
356 | 356 | |
|
|
357 | 357 | for thisProfile in profileList: |
|
|
358 | 358 | data[thisChannel,thisProfile,:] = data[thisChannel,thisProfile,:]*flip |
|
|
359 | 359 | flip *= -1.0 |
|
|
360 | 360 | |
|
|
361 | 361 | self.flip = flip |
|
|
362 | 362 | |
|
|
363 | 363 | else: |
|
|
364 | 364 | if not channelList: |
|
|
365 | 365 | data[:,:] = data[:,:]*self.flip |
|
|
366 | 366 | else: |
|
|
367 | 367 | for thisChannel in channelList: |
|
|
368 | 368 | if thisChannel not in dataOut.channelList: |
|
|
369 | 369 | continue |
|
|
370 | 370 | |
|
|
371 | 371 | data[thisChannel,:] = data[thisChannel,:]*self.flip |
|
|
372 | 372 | |
|
|
373 | 373 | self.flip *= -1. |
|
|
374 | 374 | |
|
|
375 | 375 | dataOut.data = data |
|
|
376 | 376 | |
|
|
377 | 377 | return dataOut |
|
|
378 | 378 | |
|
|
379 | 379 | |
|
|
380 | 380 | class setAttribute(Operation): |
|
|
381 | 381 | ''' |
|
|
382 | 382 | Set an arbitrary attribute(s) to dataOut |
|
|
383 | 383 | ''' |
|
|
384 | 384 | |
|
|
385 | 385 | def __init__(self): |
|
|
386 | 386 | |
|
|
387 | 387 | Operation.__init__(self) |
|
|
388 | 388 | self._ready = False |
|
|
389 | 389 | |
|
|
390 | 390 | def run(self, dataOut, **kwargs): |
|
|
391 | 391 | |
|
|
392 | 392 | for key, value in kwargs.items(): |
|
|
393 | 393 | setattr(dataOut, key, value) |
|
|
394 | 394 | |
|
|
395 | 395 | return dataOut |
|
|
396 | 396 | |
|
|
397 | 397 | |
|
|
398 | 398 | @MPDecorator |
|
|
399 | 399 | class printAttribute(Operation): |
|
|
400 | 400 | ''' |
|
|
401 | 401 | Print an arbitrary attribute of dataOut |
|
|
402 | 402 | ''' |
|
|
403 | 403 | |
|
|
404 | 404 | def __init__(self): |
|
|
405 | 405 | |
|
|
406 | 406 | Operation.__init__(self) |
|
|
407 | 407 | |
|
|
408 | 408 | def run(self, dataOut, attributes): |
|
|
409 | 409 | |
|
|
410 | 410 | if isinstance(attributes, str): |
|
|
411 | 411 | attributes = [attributes] |
|
|
412 | 412 | for attr in attributes: |
|
|
413 | 413 | if hasattr(dataOut, attr): |
|
|
414 | 414 | log.log(getattr(dataOut, attr), attr) |
|
|
415 | 415 | |
|
|
416 | 416 | |
|
|
417 | 417 | class interpolateHeights(Operation): |
|
|
418 | 418 | |
|
|
419 | 419 | def run(self, dataOut, topLim, botLim): |
|
|
420 | 420 | #69 al 72 para julia |
|
|
421 | 421 | #82-84 para meteoros |
|
|
422 | 422 | if len(numpy.shape(dataOut.data))==2: |
|
|
423 | 423 | sampInterp = (dataOut.data[:,botLim-1] + dataOut.data[:,topLim+1])/2 |
|
|
424 | 424 | sampInterp = numpy.transpose(numpy.tile(sampInterp,(topLim-botLim + 1,1))) |
|
|
425 | 425 | #dataOut.data[:,botLim:limSup+1] = sampInterp |
|
|
426 | 426 | dataOut.data[:,botLim:topLim+1] = sampInterp |
|
|
427 | 427 | else: |
|
|
428 | 428 | nHeights = dataOut.data.shape[2] |
|
|
429 | 429 | x = numpy.hstack((numpy.arange(botLim),numpy.arange(topLim+1,nHeights))) |
|
|
430 | 430 | y = dataOut.data[:,:,list(range(botLim))+list(range(topLim+1,nHeights))] |
|
|
431 | 431 | f = interpolate.interp1d(x, y, axis = 2) |
|
|
432 | 432 | xnew = numpy.arange(botLim,topLim+1) |
|
|
433 | 433 | ynew = f(xnew) |
|
|
434 | 434 | dataOut.data[:,:,botLim:topLim+1] = ynew |
|
|
435 | 435 | |
|
|
436 | 436 | return dataOut |
|
|
437 | 437 | |
|
|
438 | 438 | |
|
|
439 | 439 | class CohInt(Operation): |
|
|
440 | 440 | |
|
|
441 | 441 | isConfig = False |
|
|
442 | 442 | __profIndex = 0 |
|
|
443 | 443 | __byTime = False |
|
|
444 | 444 | __initime = None |
|
|
445 | 445 | __lastdatatime = None |
|
|
446 | 446 | __integrationtime = None |
|
|
447 | 447 | __buffer = None |
|
|
448 | 448 | __bufferStride = [] |
|
|
449 | 449 | __dataReady = False |
|
|
450 | 450 | __profIndexStride = 0 |
|
|
451 | 451 | __dataToPutStride = False |
|
|
452 | 452 | n = None |
|
|
453 | 453 | |
|
|
454 | 454 | def __init__(self, **kwargs): |
|
|
455 | 455 | |
|
|
456 | 456 | Operation.__init__(self, **kwargs) |
|
|
457 | 457 | |
|
|
458 | 458 | def setup(self, n=None, timeInterval=None, stride=None, overlapping=False, byblock=False): |
|
|
459 | 459 | """ |
|
|
460 | 460 | Set the parameters of the integration class. |
|
|
461 | 461 | |
|
|
462 | 462 | Inputs: |
|
|
463 | 463 | |
|
|
464 | 464 | n : Number of coherent integrations |
|
|
465 | 465 | timeInterval : Time of integration. If the parameter "n" is selected this one does not work |
|
|
466 | 466 | overlapping : |
|
|
467 | 467 | """ |
|
|
468 | 468 | |
|
|
469 | 469 | self.__initime = None |
|
|
470 | 470 | self.__lastdatatime = 0 |
|
|
471 | 471 | self.__buffer = None |
|
|
472 | 472 | self.__dataReady = False |
|
|
473 | 473 | self.byblock = byblock |
|
|
474 | 474 | self.stride = stride |
|
|
475 | 475 | |
|
|
476 | 476 | if n == None and timeInterval == None: |
|
|
477 | 477 | raise ValueError("n or timeInterval should be specified ...") |
|
|
478 | 478 | |
|
|
479 | 479 | if n != None: |
|
|
480 | 480 | self.n = n |
|
|
481 | 481 | self.__byTime = False |
|
|
482 | 482 | else: |
|
|
483 | 483 | self.__integrationtime = timeInterval #* 60. #if (type(timeInterval)!=integer) -> change this line |
|
|
484 | 484 | self.n = 9999 |
|
|
485 | 485 | self.__byTime = True |
|
|
486 | 486 | |
|
|
487 | 487 | if overlapping: |
|
|
488 | 488 | self.__withOverlapping = True |
|
|
489 | 489 | self.__buffer = None |
|
|
490 | 490 | else: |
|
|
491 | 491 | self.__withOverlapping = False |
|
|
492 | 492 | self.__buffer = 0 |
|
|
493 | 493 | |
|
|
494 | 494 | self.__profIndex = 0 |
|
|
495 | 495 | |
|
|
496 | 496 | def putData(self, data): |
|
|
497 | 497 | |
|
|
498 | 498 | """ |
|
|
499 | 499 | Add a profile to the __buffer and increase in one the __profileIndex |
|
|
500 | 500 | |
|
|
501 | 501 | """ |
|
|
502 | 502 | |
|
|
503 | 503 | if not self.__withOverlapping: |
|
|
504 | 504 | self.__buffer += data.copy() |
|
|
505 | 505 | self.__profIndex += 1 |
|
|
506 | 506 | return |
|
|
507 | 507 | |
|
|
508 | 508 | #Overlapping data |
|
|
509 | 509 | nChannels, nHeis = data.shape |
|
|
510 | 510 | data = numpy.reshape(data, (1, nChannels, nHeis)) |
|
|
511 | 511 | |
|
|
512 | 512 | #If the buffer is empty then it takes the data value |
|
|
513 | 513 | if self.__buffer is None: |
|
|
514 | 514 | self.__buffer = data |
|
|
515 | 515 | self.__profIndex += 1 |
|
|
516 | 516 | return |
|
|
517 | 517 | |
|
|
518 | 518 | #If the buffer length is lower than n then stakcing the data value |
|
|
519 | 519 | if self.__profIndex < self.n: |
|
|
520 | 520 | self.__buffer = numpy.vstack((self.__buffer, data)) |
|
|
521 | 521 | self.__profIndex += 1 |
|
|
522 | 522 | return |
|
|
523 | 523 | |
|
|
524 | 524 | #If the buffer length is equal to n then replacing the last buffer value with the data value |
|
|
525 | 525 | self.__buffer = numpy.roll(self.__buffer, -1, axis=0) |
|
|
526 | 526 | self.__buffer[self.n-1] = data |
|
|
527 | 527 | self.__profIndex = self.n |
|
|
528 | 528 | return |
|
|
529 | 529 | |
|
|
530 | 530 | |
|
|
531 | 531 | def pushData(self): |
|
|
532 | 532 | """ |
|
|
533 | 533 | Return the sum of the last profiles and the profiles used in the sum. |
|
|
534 | 534 | |
|
|
535 | 535 | Affected: |
|
|
536 | 536 | |
|
|
537 | 537 | self.__profileIndex |
|
|
538 | 538 | |
|
|
539 | 539 | """ |
|
|
540 | 540 | |
|
|
541 | 541 | if not self.__withOverlapping: |
|
|
542 | 542 | data = self.__buffer |
|
|
543 | 543 | n = self.__profIndex |
|
|
544 | 544 | |
|
|
545 | 545 | self.__buffer = 0 |
|
|
546 | 546 | self.__profIndex = 0 |
|
|
547 | 547 | |
|
|
548 | 548 | return data, n |
|
|
549 | 549 | |
|
|
550 | 550 | #Integration with Overlapping |
|
|
551 | 551 | data = numpy.sum(self.__buffer, axis=0) |
|
|
552 | 552 | # print data |
|
|
553 | 553 | # raise |
|
|
554 | 554 | n = self.__profIndex |
|
|
555 | 555 | |
|
|
556 | 556 | return data, n |
|
|
557 | 557 | |
|
|
558 | 558 | def byProfiles(self, data): |
|
|
559 | 559 | |
|
|
560 | 560 | self.__dataReady = False |
|
|
561 | 561 | avgdata = None |
|
|
562 | 562 | # n = None |
|
|
563 | 563 | # print data |
|
|
564 | 564 | # raise |
|
|
565 | 565 | self.putData(data) |
|
|
566 | 566 | |
|
|
567 | 567 | if self.__profIndex == self.n: |
|
|
568 | 568 | avgdata, n = self.pushData() |
|
|
569 | 569 | self.__dataReady = True |
|
|
570 | 570 | |
|
|
571 | 571 | return avgdata |
|
|
572 | 572 | |
|
|
573 | 573 | def byTime(self, data, datatime): |
|
|
574 | 574 | |
|
|
575 | 575 | self.__dataReady = False |
|
|
576 | 576 | avgdata = None |
|
|
577 | 577 | n = None |
|
|
578 | 578 | |
|
|
579 | 579 | self.putData(data) |
|
|
580 | 580 | |
|
|
581 | 581 | if (datatime - self.__initime) >= self.__integrationtime: |
|
|
582 | 582 | avgdata, n = self.pushData() |
|
|
583 | 583 | self.n = n |
|
|
584 | 584 | self.__dataReady = True |
|
|
585 | 585 | |
|
|
586 | 586 | return avgdata |
|
|
587 | 587 | |
|
|
588 | 588 | def integrateByStride(self, data, datatime): |
|
|
589 | 589 | # print data |
|
|
590 | 590 | if self.__profIndex == 0: |
|
|
591 | 591 | self.__buffer = [[data.copy(), datatime]] |
|
|
592 | 592 | else: |
|
|
593 | 593 | self.__buffer.append([data.copy(),datatime]) |
|
|
594 | 594 | self.__profIndex += 1 |
|
|
595 | 595 | self.__dataReady = False |
|
|
596 | 596 | |
|
|
597 | 597 | if self.__profIndex == self.n * self.stride : |
|
|
598 | 598 | self.__dataToPutStride = True |
|
|
599 | 599 | self.__profIndexStride = 0 |
|
|
600 | 600 | self.__profIndex = 0 |
|
|
601 | 601 | self.__bufferStride = [] |
|
|
602 | 602 | for i in range(self.stride): |
|
|
603 | 603 | current = self.__buffer[i::self.stride] |
|
|
604 | 604 | data = numpy.sum([t[0] for t in current], axis=0) |
|
|
605 | 605 | avgdatatime = numpy.average([t[1] for t in current]) |
|
|
606 | 606 | # print data |
|
|
607 | 607 | self.__bufferStride.append((data, avgdatatime)) |
|
|
608 | 608 | |
|
|
609 | 609 | if self.__dataToPutStride: |
|
|
610 | 610 | self.__dataReady = True |
|
|
611 | 611 | self.__profIndexStride += 1 |
|
|
612 | 612 | if self.__profIndexStride == self.stride: |
|
|
613 | 613 | self.__dataToPutStride = False |
|
|
614 | 614 | # print self.__bufferStride[self.__profIndexStride - 1] |
|
|
615 | 615 | # raise |
|
|
616 | 616 | return self.__bufferStride[self.__profIndexStride - 1] |
|
|
617 | 617 | |
|
|
618 | 618 | |
|
|
619 | 619 | return None, None |
|
|
620 | 620 | |
|
|
621 | 621 | def integrate(self, data, datatime=None): |
|
|
622 | 622 | |
|
|
623 | 623 | if self.__initime == None: |
|
|
624 | 624 | self.__initime = datatime |
|
|
625 | 625 | |
|
|
626 | 626 | if self.__byTime: |
|
|
627 | 627 | avgdata = self.byTime(data, datatime) |
|
|
628 | 628 | else: |
|
|
629 | 629 | avgdata = self.byProfiles(data) |
|
|
630 | 630 | |
|
|
631 | 631 | |
|
|
632 | 632 | self.__lastdatatime = datatime |
|
|
633 | 633 | |
|
|
634 | 634 | if avgdata is None: |
|
|
635 | 635 | return None, None |
|
|
636 | 636 | |
|
|
637 | 637 | avgdatatime = self.__initime |
|
|
638 | 638 | |
|
|
639 | 639 | deltatime = datatime - self.__lastdatatime |
|
|
640 | 640 | |
|
|
641 | 641 | if not self.__withOverlapping: |
|
|
642 | 642 | self.__initime = datatime |
|
|
643 | 643 | else: |
|
|
644 | 644 | self.__initime += deltatime |
|
|
645 | 645 | |
|
|
646 | 646 | return avgdata, avgdatatime |
|
|
647 | 647 | |
|
|
648 | 648 | def integrateByBlock(self, dataOut): |
|
|
649 | 649 | |
|
|
650 | 650 | times = int(dataOut.data.shape[1]/self.n) |
|
|
651 | 651 | avgdata = numpy.zeros((dataOut.nChannels, times, dataOut.nHeights), dtype=numpy.complex) |
|
|
652 | 652 | |
|
|
653 | 653 | id_min = 0 |
|
|
654 | 654 | id_max = self.n |
|
|
655 | 655 | |
|
|
656 | 656 | for i in range(times): |
|
|
657 | 657 | junk = dataOut.data[:,id_min:id_max,:] |
|
|
658 | 658 | avgdata[:,i,:] = junk.sum(axis=1) |
|
|
659 | 659 | id_min += self.n |
|
|
660 | 660 | id_max += self.n |
|
|
661 | 661 | |
|
|
662 | 662 | timeInterval = dataOut.ippSeconds*self.n |
|
|
663 | 663 | avgdatatime = (times - 1) * timeInterval + dataOut.utctime |
|
|
664 | 664 | self.__dataReady = True |
|
|
665 | 665 | return avgdata, avgdatatime |
|
|
666 | 666 | |
|
|
667 | 667 | def run(self, dataOut, n=None, timeInterval=None, stride=None, overlapping=False, byblock=False, **kwargs): |
|
|
668 | 668 | |
|
|
669 | 669 | if not self.isConfig: |
|
|
670 | 670 | self.setup(n=n, stride=stride, timeInterval=timeInterval, overlapping=overlapping, byblock=byblock, **kwargs) |
|
|
671 | 671 | self.isConfig = True |
|
|
672 | 672 | |
|
|
673 | 673 | if dataOut.flagDataAsBlock: |
|
|
674 | 674 | """ |
|
|
675 | 675 | Si la data es leida por bloques, dimension = [nChannels, nProfiles, nHeis] |
|
|
676 | 676 | """ |
|
|
677 | 677 | avgdata, avgdatatime = self.integrateByBlock(dataOut) |
|
|
678 | 678 | dataOut.nProfiles /= self.n |
|
|
679 | 679 | else: |
|
|
680 | 680 | if stride is None: |
|
|
681 | 681 | avgdata, avgdatatime = self.integrate(dataOut.data, dataOut.utctime) |
|
|
682 | 682 | else: |
|
|
683 | 683 | avgdata, avgdatatime = self.integrateByStride(dataOut.data, dataOut.utctime) |
|
|
684 | 684 | |
|
|
685 | 685 | |
|
|
686 | 686 | # dataOut.timeInterval *= n |
|
|
687 | 687 | dataOut.flagNoData = True |
|
|
688 | 688 | |
|
|
689 | 689 | if self.__dataReady: |
|
|
690 | 690 | dataOut.data = avgdata |
|
|
691 | 691 | if not dataOut.flagCohInt: |
|
|
692 | 692 | dataOut.nCohInt *= self.n |
|
|
693 | 693 | dataOut.flagCohInt = True |
|
|
694 | 694 | dataOut.utctime = avgdatatime |
|
|
695 | 695 | # print avgdata, avgdatatime |
|
|
696 | 696 | # raise |
|
|
697 | 697 | # dataOut.timeInterval = dataOut.ippSeconds * dataOut.nCohInt |
|
|
698 | 698 | dataOut.flagNoData = False |
|
|
699 | 699 | return dataOut |
|
|
700 | 700 | |
|
|
701 | 701 | class Decoder(Operation): |
|
|
702 | 702 | |
|
|
703 | 703 | isConfig = False |
|
|
704 | 704 | __profIndex = 0 |
|
|
705 | 705 | |
|
|
706 | 706 | code = None |
|
|
707 | 707 | |
|
|
708 | 708 | nCode = None |
|
|
709 | 709 | nBaud = None |
|
|
710 | 710 | |
|
|
711 | 711 | def __init__(self, **kwargs): |
|
|
712 | 712 | |
|
|
713 | 713 | Operation.__init__(self, **kwargs) |
|
|
714 | 714 | |
|
|
715 | 715 | self.times = None |
|
|
716 | 716 | self.osamp = None |
|
|
717 | 717 | # self.__setValues = False |
|
|
718 | 718 | self.isConfig = False |
|
|
719 | 719 | self.setupReq = False |
|
|
720 | 720 | def setup(self, code, osamp, dataOut): |
|
|
721 | 721 | |
|
|
722 | 722 | self.__profIndex = 0 |
|
|
723 | 723 | |
|
|
724 | 724 | self.code = code |
|
|
725 | 725 | |
|
|
726 | 726 | self.nCode = len(code) |
|
|
727 | 727 | self.nBaud = len(code[0]) |
|
|
728 | 728 | if (osamp != None) and (osamp >1): |
|
|
729 | 729 | self.osamp = osamp |
|
|
730 | 730 | self.code = numpy.repeat(code, repeats=self.osamp, axis=1) |
|
|
731 | 731 | self.nBaud = self.nBaud*self.osamp |
|
|
732 | 732 | |
|
|
733 | 733 | self.__nChannels = dataOut.nChannels |
|
|
734 | 734 | self.__nProfiles = dataOut.nProfiles |
|
|
735 | 735 | self.__nHeis = dataOut.nHeights |
|
|
736 | 736 | |
|
|
737 | 737 | if self.__nHeis < self.nBaud: |
|
|
738 | 738 | raise ValueError('Number of heights (%d) should be greater than number of bauds (%d)' %(self.__nHeis, self.nBaud)) |
|
|
739 | 739 | |
|
|
740 | 740 | #Frequency |
|
|
741 | 741 | __codeBuffer = numpy.zeros((self.nCode, self.__nHeis), dtype=numpy.complex) |
|
|
742 | 742 | |
|
|
743 | 743 | __codeBuffer[:,0:self.nBaud] = self.code |
|
|
744 | 744 | |
|
|
745 | 745 | self.fft_code = numpy.conj(numpy.fft.fft(__codeBuffer, axis=1)) |
|
|
746 | 746 | |
|
|
747 | 747 | if dataOut.flagDataAsBlock: |
|
|
748 | 748 | |
|
|
749 | 749 | self.ndatadec = self.__nHeis #- self.nBaud + 1 |
|
|
750 | 750 | |
|
|
751 | 751 | self.datadecTime = numpy.zeros((self.__nChannels, self.__nProfiles, self.ndatadec), dtype=numpy.complex) |
|
|
752 | 752 | |
|
|
753 | 753 | else: |
|
|
754 | 754 | |
|
|
755 | 755 | #Time |
|
|
756 | 756 | self.ndatadec = self.__nHeis #- self.nBaud + 1 |
|
|
757 | 757 | |
|
|
758 | 758 | self.datadecTime = numpy.zeros((self.__nChannels, self.ndatadec), dtype=numpy.complex) |
|
|
759 | 759 | |
|
|
760 | 760 | def __convolutionInFreq(self, data): |
|
|
761 | 761 | |
|
|
762 | 762 | fft_code = self.fft_code[self.__profIndex].reshape(1,-1) |
|
|
763 | 763 | |
|
|
764 | 764 | fft_data = numpy.fft.fft(data, axis=1) |
|
|
765 | 765 | |
|
|
766 | 766 | conv = fft_data*fft_code |
|
|
767 | 767 | |
|
|
768 | 768 | data = numpy.fft.ifft(conv,axis=1) |
|
|
769 | 769 | |
|
|
770 | 770 | return data |
|
|
771 | 771 | |
|
|
772 | 772 | def __convolutionInFreqOpt(self, data): |
|
|
773 | 773 | |
|
|
774 | 774 | raise NotImplementedError |
|
|
775 | 775 | |
|
|
776 | 776 | def __convolutionInTime(self, data): |
|
|
777 | 777 | |
|
|
778 | 778 | code = self.code[self.__profIndex] |
|
|
779 | 779 | for i in range(self.__nChannels): |
|
|
780 | 780 | self.datadecTime[i,:] = numpy.correlate(data[i,:], code, mode='full')[self.nBaud-1:] |
|
|
781 | 781 | |
|
|
782 | 782 | return self.datadecTime |
|
|
783 | 783 | |
|
|
784 | 784 | def __convolutionByBlockInTime(self, data): |
|
|
785 | 785 | |
|
|
786 | 786 | repetitions = int(self.__nProfiles / self.nCode) |
|
|
787 | 787 | junk = numpy.lib.stride_tricks.as_strided(self.code, (repetitions, self.code.size), (0, self.code.itemsize)) |
|
|
788 | 788 | junk = junk.flatten() |
|
|
789 | 789 | code_block = numpy.reshape(junk, (self.nCode*repetitions, self.nBaud)) |
|
|
790 | 790 | profilesList = range(self.__nProfiles) |
|
|
791 | 791 | |
|
|
792 | 792 | for i in range(self.__nChannels): |
|
|
793 | 793 | for j in profilesList: |
|
|
794 | 794 | self.datadecTime[i,j,:] = numpy.correlate(data[i,j,:], code_block[j,:], mode='full')[self.nBaud-1:] |
|
|
795 | 795 | return self.datadecTime |
|
|
796 | 796 | |
|
|
797 | 797 | def __convolutionByBlockInFreq(self, data): |
|
|
798 | 798 | |
|
|
799 | 799 | raise NotImplementedError("Decoder by frequency fro Blocks not implemented") |
|
|
800 | 800 | |
|
|
801 | 801 | |
|
|
802 | 802 | fft_code = self.fft_code[self.__profIndex].reshape(1,-1) |
|
|
803 | 803 | |
|
|
804 | 804 | fft_data = numpy.fft.fft(data, axis=2) |
|
|
805 | 805 | |
|
|
806 | 806 | conv = fft_data*fft_code |
|
|
807 | 807 | |
|
|
808 | 808 | data = numpy.fft.ifft(conv,axis=2) |
|
|
809 | 809 | |
|
|
810 | 810 | return data |
|
|
811 | 811 | |
|
|
812 | 812 | |
|
|
813 | 813 | def run(self, dataOut, code=None, nCode=None, nBaud=None, mode = 0, osamp=None, times=None): |
|
|
814 | 814 | |
|
|
815 | 815 | if dataOut.flagDecodeData: |
|
|
816 | 816 | print("This data is already decoded, recoding again ...") |
|
|
817 | 817 | |
|
|
818 | 818 | if not self.isConfig: |
|
|
819 | 819 | |
|
|
820 | 820 | if code is None: |
|
|
821 | 821 | if dataOut.code is None: |
|
|
822 | 822 | raise ValueError("Code could not be read from %s instance. Enter a value in Code parameter" %dataOut.type) |
|
|
823 | 823 | |
|
|
824 | 824 | code = dataOut.code |
|
|
825 | 825 | else: |
|
|
826 | 826 | code = numpy.array(code).reshape(nCode,nBaud) |
|
|
827 | 827 | self.setup(code, osamp, dataOut) |
|
|
828 | 828 | |
|
|
829 | 829 | self.isConfig = True |
|
|
830 | 830 | |
|
|
831 | 831 | if mode == 3: |
|
|
832 | 832 | sys.stderr.write("Decoder Warning: mode=%d is not valid, using mode=0\n" %mode) |
|
|
833 | 833 | |
|
|
834 | 834 | if times != None: |
|
|
835 | 835 | sys.stderr.write("Decoder Warning: Argument 'times' in not used anymore\n") |
|
|
836 | 836 | |
|
|
837 | 837 | if self.code is None: |
|
|
838 | 838 | print("Fail decoding: Code is not defined.") |
|
|
839 | 839 | return |
|
|
840 | 840 | |
|
|
841 | 841 | self.__nProfiles = dataOut.nProfiles |
|
|
842 | 842 | datadec = None |
|
|
843 | 843 | |
|
|
844 | 844 | if mode == 3: |
|
|
845 | 845 | mode = 0 |
|
|
846 | 846 | |
|
|
847 | 847 | if dataOut.flagDataAsBlock: |
|
|
848 | 848 | """ |
|
|
849 | 849 | Decoding when data have been read as block, |
|
|
850 | 850 | """ |
|
|
851 | 851 | |
|
|
852 | 852 | if mode == 0: |
|
|
853 | 853 | datadec = self.__convolutionByBlockInTime(dataOut.data) |
|
|
854 | 854 | if mode == 1: |
|
|
855 | 855 | datadec = self.__convolutionByBlockInFreq(dataOut.data) |
|
|
856 | 856 | else: |
|
|
857 | 857 | """ |
|
|
858 | 858 | Decoding when data have been read profile by profile |
|
|
859 | 859 | """ |
|
|
860 | 860 | if mode == 0: |
|
|
861 | 861 | datadec = self.__convolutionInTime(dataOut.data) |
|
|
862 | 862 | |
|
|
863 | 863 | if mode == 1: |
|
|
864 | 864 | datadec = self.__convolutionInFreq(dataOut.data) |
|
|
865 | 865 | |
|
|
866 | 866 | if mode == 2: |
|
|
867 | 867 | datadec = self.__convolutionInFreqOpt(dataOut.data) |
|
|
868 | 868 | |
|
|
869 | 869 | if datadec is None: |
|
|
870 | 870 | raise ValueError("Codification mode selected is not valid: mode=%d. Try selecting 0 or 1" %mode) |
|
|
871 | 871 | |
|
|
872 | 872 | dataOut.code = self.code |
|
|
873 | 873 | dataOut.nCode = self.nCode |
|
|
874 | 874 | dataOut.nBaud = self.nBaud |
|
|
875 | 875 | |
|
|
876 | 876 | dataOut.data = datadec |
|
|
877 | 877 | |
|
|
878 | 878 | dataOut.heightList = dataOut.heightList[0:datadec.shape[-1]] |
|
|
879 | 879 | |
|
|
880 | 880 | dataOut.flagDecodeData = True #asumo q la data esta decodificada |
|
|
881 | 881 | |
|
|
882 | 882 | if self.__profIndex == self.nCode-1: |
|
|
883 | 883 | self.__profIndex = 0 |
|
|
884 | 884 | return dataOut |
|
|
885 | 885 | |
|
|
886 | 886 | self.__profIndex += 1 |
|
|
887 | 887 | |
|
|
888 | 888 | return dataOut |
|
|
889 | 889 | # dataOut.flagDeflipData = True #asumo q la data no esta sin flip |
|
|
890 | 890 | |
|
|
891 | 891 | |
|
|
892 | 892 | class ProfileConcat(Operation): |
|
|
893 | 893 | |
|
|
894 | 894 | isConfig = False |
|
|
895 | 895 | buffer = None |
|
|
896 | 896 | |
|
|
897 | 897 | def __init__(self, **kwargs): |
|
|
898 | 898 | |
|
|
899 | 899 | Operation.__init__(self, **kwargs) |
|
|
900 | 900 | self.profileIndex = 0 |
|
|
901 | 901 | |
|
|
902 | 902 | def reset(self): |
|
|
903 | 903 | self.buffer = numpy.zeros_like(self.buffer) |
|
|
904 | 904 | self.start_index = 0 |
|
|
905 | 905 | self.times = 1 |
|
|
906 | 906 | |
|
|
907 | 907 | def setup(self, data, m, n=1): |
|
|
908 | 908 | self.buffer = numpy.zeros((data.shape[0],data.shape[1]*m),dtype=type(data[0,0])) |
|
|
909 | 909 | self.nHeights = data.shape[1]#.nHeights |
|
|
910 | 910 | self.start_index = 0 |
|
|
911 | 911 | self.times = 1 |
|
|
912 | 912 | |
|
|
913 | 913 | def concat(self, data): |
|
|
914 | 914 | |
|
|
915 | 915 | self.buffer[:,self.start_index:self.nHeights*self.times] = data.copy() |
|
|
916 | 916 | self.start_index = self.start_index + self.nHeights |
|
|
917 | 917 | |
|
|
918 | 918 | def run(self, dataOut, m): |
|
|
919 | 919 | dataOut.flagNoData = True |
|
|
920 | 920 | |
|
|
921 | 921 | if not self.isConfig: |
|
|
922 | 922 | self.setup(dataOut.data, m, 1) |
|
|
923 | 923 | self.isConfig = True |
|
|
924 | 924 | |
|
|
925 | 925 | if dataOut.flagDataAsBlock: |
|
|
926 | 926 | raise ValueError("ProfileConcat can only be used when voltage have been read profile by profile, getBlock = False") |
|
|
927 | 927 | |
|
|
928 | 928 | else: |
|
|
929 | 929 | self.concat(dataOut.data) |
|
|
930 | 930 | self.times += 1 |
|
|
931 | 931 | if self.times > m: |
|
|
932 | 932 | dataOut.data = self.buffer |
|
|
933 | 933 | self.reset() |
|
|
934 | 934 | dataOut.flagNoData = False |
|
|
935 | 935 | # se deben actualizar mas propiedades del header y del objeto dataOut, por ejemplo, las alturas |
|
|
936 | 936 | deltaHeight = dataOut.heightList[1] - dataOut.heightList[0] |
|
|
937 | 937 | xf = dataOut.heightList[0] + dataOut.nHeights * deltaHeight * m |
|
|
938 | 938 | dataOut.heightList = numpy.arange(dataOut.heightList[0], xf, deltaHeight) |
|
|
939 | 939 | dataOut.ippSeconds *= m |
|
|
940 | 940 | return dataOut |
|
|
941 | 941 | |
|
|
942 | 942 | class ProfileSelector(Operation): |
|
|
943 | 943 | |
|
|
944 | 944 | profileIndex = None |
|
|
945 | 945 | # Tamanho total de los perfiles |
|
|
946 | 946 | nProfiles = None |
|
|
947 | 947 | |
|
|
948 | 948 | def __init__(self, **kwargs): |
|
|
949 | 949 | |
|
|
950 | 950 | Operation.__init__(self, **kwargs) |
|
|
951 | 951 | self.profileIndex = 0 |
|
|
952 | 952 | |
|
|
953 | 953 | def incProfileIndex(self): |
|
|
954 | 954 | |
|
|
955 | 955 | self.profileIndex += 1 |
|
|
956 | 956 | |
|
|
957 | 957 | if self.profileIndex >= self.nProfiles: |
|
|
958 | 958 | self.profileIndex = 0 |
|
|
959 | 959 | |
|
|
960 | 960 | def isThisProfileInRange(self, profileIndex, minIndex, maxIndex): |
|
|
961 | 961 | |
|
|
962 | 962 | if profileIndex < minIndex: |
|
|
963 | 963 | return False |
|
|
964 | 964 | |
|
|
965 | 965 | if profileIndex > maxIndex: |
|
|
966 | 966 | return False |
|
|
967 | 967 | |
|
|
968 | 968 | return True |
|
|
969 | 969 | |
|
|
970 | 970 | def isThisProfileInList(self, profileIndex, profileList): |
|
|
971 | 971 | |
|
|
972 | 972 | if profileIndex not in profileList: |
|
|
973 | 973 | return False |
|
|
974 | 974 | |
|
|
975 | 975 | return True |
|
|
976 | 976 | |
|
|
977 | 977 | def run(self, dataOut, profileList=None, profileRangeList=None, beam=None, byblock=False, rangeList = None, nProfiles=None): |
|
|
978 | 978 | |
|
|
979 | 979 | """ |
|
|
980 | 980 | ProfileSelector: |
|
|
981 | 981 | |
|
|
982 | 982 | Inputs: |
|
|
983 | 983 | profileList : Index of profiles selected. Example: profileList = (0,1,2,7,8) |
|
|
984 | 984 | |
|
|
985 | 985 | profileRangeList : Minimum and maximum profile indexes. Example: profileRangeList = (4, 30) |
|
|
986 | 986 | |
|
|
987 | 987 | rangeList : List of profile ranges. Example: rangeList = ((4, 30), (32, 64), (128, 256)) |
|
|
988 | 988 | |
|
|
989 | 989 | """ |
|
|
990 | 990 | |
|
|
991 | 991 | if rangeList is not None: |
|
|
992 | 992 | if type(rangeList[0]) not in (tuple, list): |
|
|
993 | 993 | rangeList = [rangeList] |
|
|
994 | 994 | |
|
|
995 | 995 | dataOut.flagNoData = True |
|
|
996 | 996 | |
|
|
997 | 997 | if dataOut.flagDataAsBlock: |
|
|
998 | 998 | """ |
|
|
999 | 999 | data dimension = [nChannels, nProfiles, nHeis] |
|
|
1000 | 1000 | """ |
|
|
1001 | 1001 | if profileList != None: |
|
|
1002 | 1002 | dataOut.data = dataOut.data[:,profileList,:] |
|
|
1003 | 1003 | |
|
|
1004 | 1004 | if profileRangeList != None: |
|
|
1005 | 1005 | minIndex = profileRangeList[0] |
|
|
1006 | 1006 | maxIndex = profileRangeList[1] |
|
|
1007 | 1007 | profileList = list(range(minIndex, maxIndex+1)) |
|
|
1008 | 1008 | |
|
|
1009 | 1009 | dataOut.data = dataOut.data[:,minIndex:maxIndex+1,:] |
|
|
1010 | 1010 | |
|
|
1011 | 1011 | if rangeList != None: |
|
|
1012 | 1012 | |
|
|
1013 | 1013 | profileList = [] |
|
|
1014 | 1014 | |
|
|
1015 | 1015 | for thisRange in rangeList: |
|
|
1016 | 1016 | minIndex = thisRange[0] |
|
|
1017 | 1017 | maxIndex = thisRange[1] |
|
|
1018 | 1018 | |
|
|
1019 | 1019 | profileList.extend(list(range(minIndex, maxIndex+1))) |
|
|
1020 | 1020 | |
|
|
1021 | 1021 | dataOut.data = dataOut.data[:,profileList,:] |
|
|
1022 | 1022 | |
|
|
1023 | 1023 | dataOut.nProfiles = len(profileList) |
|
|
1024 | 1024 | dataOut.profileIndex = dataOut.nProfiles - 1 |
|
|
1025 | 1025 | dataOut.flagNoData = False |
|
|
1026 | 1026 | |
|
|
1027 | 1027 | return dataOut |
|
|
1028 | 1028 | |
|
|
1029 | 1029 | """ |
|
|
1030 | 1030 | data dimension = [nChannels, nHeis] |
|
|
1031 | 1031 | """ |
|
|
1032 | 1032 | |
|
|
1033 | 1033 | if profileList != None: |
|
|
1034 | 1034 | |
|
|
1035 | 1035 | if self.isThisProfileInList(dataOut.profileIndex, profileList): |
|
|
1036 | 1036 | |
|
|
1037 | 1037 | self.nProfiles = len(profileList) |
|
|
1038 | 1038 | dataOut.nProfiles = self.nProfiles |
|
|
1039 | 1039 | dataOut.profileIndex = self.profileIndex |
|
|
1040 | 1040 | dataOut.flagNoData = False |
|
|
1041 | 1041 | |
|
|
1042 | 1042 | self.incProfileIndex() |
|
|
1043 | 1043 | return dataOut |
|
|
1044 | 1044 | |
|
|
1045 | 1045 | if profileRangeList != None: |
|
|
1046 | 1046 | |
|
|
1047 | 1047 | minIndex = profileRangeList[0] |
|
|
1048 | 1048 | maxIndex = profileRangeList[1] |
|
|
1049 | 1049 | |
|
|
1050 | 1050 | if self.isThisProfileInRange(dataOut.profileIndex, minIndex, maxIndex): |
|
|
1051 | 1051 | |
|
|
1052 | 1052 | self.nProfiles = maxIndex - minIndex + 1 |
|
|
1053 | 1053 | dataOut.nProfiles = self.nProfiles |
|
|
1054 | 1054 | dataOut.profileIndex = self.profileIndex |
|
|
1055 | 1055 | dataOut.flagNoData = False |
|
|
1056 | 1056 | |
|
|
1057 | 1057 | self.incProfileIndex() |
|
|
1058 | 1058 | return dataOut |
|
|
1059 | 1059 | |
|
|
1060 | 1060 | if rangeList != None: |
|
|
1061 | 1061 | |
|
|
1062 | 1062 | nProfiles = 0 |
|
|
1063 | 1063 | |
|
|
1064 | 1064 | for thisRange in rangeList: |
|
|
1065 | 1065 | minIndex = thisRange[0] |
|
|
1066 | 1066 | maxIndex = thisRange[1] |
|
|
1067 | 1067 | |
|
|
1068 | 1068 | nProfiles += maxIndex - minIndex + 1 |
|
|
1069 | 1069 | |
|
|
1070 | 1070 | for thisRange in rangeList: |
|
|
1071 | 1071 | |
|
|
1072 | 1072 | minIndex = thisRange[0] |
|
|
1073 | 1073 | maxIndex = thisRange[1] |
|
|
1074 | 1074 | |
|
|
1075 | 1075 | if self.isThisProfileInRange(dataOut.profileIndex, minIndex, maxIndex): |
|
|
1076 | 1076 | |
|
|
1077 | 1077 | self.nProfiles = nProfiles |
|
|
1078 | 1078 | dataOut.nProfiles = self.nProfiles |
|
|
1079 | 1079 | dataOut.profileIndex = self.profileIndex |
|
|
1080 | 1080 | dataOut.flagNoData = False |
|
|
1081 | 1081 | |
|
|
1082 | 1082 | self.incProfileIndex() |
|
|
1083 | 1083 | |
|
|
1084 | 1084 | break |
|
|
1085 | 1085 | |
|
|
1086 | 1086 | return dataOut |
|
|
1087 | 1087 | |
|
|
1088 | 1088 | |
|
|
1089 | 1089 | if beam != None: #beam is only for AMISR data |
|
|
1090 | 1090 | if self.isThisProfileInList(dataOut.profileIndex, dataOut.beamRangeDict[beam]): |
|
|
1091 | 1091 | dataOut.flagNoData = False |
|
|
1092 | 1092 | dataOut.profileIndex = self.profileIndex |
|
|
1093 | 1093 | |
|
|
1094 | 1094 | self.incProfileIndex() |
|
|
1095 | 1095 | |
|
|
1096 | 1096 | return dataOut |
|
|
1097 | 1097 | |
|
|
1098 | 1098 | raise ValueError("ProfileSelector needs profileList, profileRangeList or rangeList parameter") |
|
|
1099 | 1099 | |
|
|
1100 | 1100 | |
|
|
1101 | 1101 | class Reshaper(Operation): |
|
|
1102 | 1102 | |
|
|
1103 | 1103 | def __init__(self, **kwargs): |
|
|
1104 | 1104 | |
|
|
1105 | 1105 | Operation.__init__(self, **kwargs) |
|
|
1106 | 1106 | |
|
|
1107 | 1107 | self.__buffer = None |
|
|
1108 | 1108 | self.__nitems = 0 |
|
|
1109 | 1109 | |
|
|
1110 | 1110 | def __appendProfile(self, dataOut, nTxs): |
|
|
1111 | 1111 | |
|
|
1112 | 1112 | if self.__buffer is None: |
|
|
1113 | 1113 | shape = (dataOut.nChannels, int(dataOut.nHeights/nTxs) ) |
|
|
1114 | 1114 | self.__buffer = numpy.empty(shape, dtype = dataOut.data.dtype) |
|
|
1115 | 1115 | |
|
|
1116 | 1116 | ini = dataOut.nHeights * self.__nitems |
|
|
1117 | 1117 | end = ini + dataOut.nHeights |
|
|
1118 | 1118 | |
|
|
1119 | 1119 | self.__buffer[:, ini:end] = dataOut.data |
|
|
1120 | 1120 | |
|
|
1121 | 1121 | self.__nitems += 1 |
|
|
1122 | 1122 | |
|
|
1123 | 1123 | return int(self.__nitems*nTxs) |
|
|
1124 | 1124 | |
|
|
1125 | 1125 | def __getBuffer(self): |
|
|
1126 | 1126 | |
|
|
1127 | 1127 | if self.__nitems == int(1./self.__nTxs): |
|
|
1128 | 1128 | |
|
|
1129 | 1129 | self.__nitems = 0 |
|
|
1130 | 1130 | |
|
|
1131 | 1131 | return self.__buffer.copy() |
|
|
1132 | 1132 | |
|
|
1133 | 1133 | return None |
|
|
1134 | 1134 | |
|
|
1135 | 1135 | def __checkInputs(self, dataOut, shape, nTxs): |
|
|
1136 | 1136 | |
|
|
1137 | 1137 | if shape is None and nTxs is None: |
|
|
1138 | 1138 | raise ValueError("Reshaper: shape of factor should be defined") |
|
|
1139 | 1139 | |
|
|
1140 | 1140 | if nTxs: |
|
|
1141 | 1141 | if nTxs < 0: |
|
|
1142 | 1142 | raise ValueError("nTxs should be greater than 0") |
|
|
1143 | 1143 | |
|
|
1144 | 1144 | if nTxs < 1 and dataOut.nProfiles % (1./nTxs) != 0: |
|
|
1145 | 1145 | raise ValueError("nProfiles= %d is not divisibled by (1./nTxs) = %f" %(dataOut.nProfiles, (1./nTxs))) |
|
|
1146 | 1146 | |
|
|
1147 | 1147 | shape = [dataOut.nChannels, dataOut.nProfiles*nTxs, dataOut.nHeights/nTxs] |
|
|
1148 | 1148 | |
|
|
1149 | 1149 | return shape, nTxs |
|
|
1150 | 1150 | |
|
|
1151 | 1151 | if len(shape) != 2 and len(shape) != 3: |
|
|
1152 | 1152 | raise ValueError("shape dimension should be equal to 2 or 3. shape = (nProfiles, nHeis) or (nChannels, nProfiles, nHeis). Actually shape = (%d, %d, %d)" %(dataOut.nChannels, dataOut.nProfiles, dataOut.nHeights)) |
|
|
1153 | 1153 | |
|
|
1154 | 1154 | if len(shape) == 2: |
|
|
1155 | 1155 | shape_tuple = [dataOut.nChannels] |
|
|
1156 | 1156 | shape_tuple.extend(shape) |
|
|
1157 | 1157 | else: |
|
|
1158 | 1158 | shape_tuple = list(shape) |
|
|
1159 | 1159 | |
|
|
1160 | 1160 | nTxs = 1.0*shape_tuple[1]/dataOut.nProfiles |
|
|
1161 | 1161 | |
|
|
1162 | 1162 | return shape_tuple, nTxs |
|
|
1163 | 1163 | |
|
|
1164 | 1164 | def run(self, dataOut, shape=None, nTxs=None): |
|
|
1165 | 1165 | |
|
|
1166 | 1166 | shape_tuple, self.__nTxs = self.__checkInputs(dataOut, shape, nTxs) |
|
|
1167 | 1167 | |
|
|
1168 | 1168 | dataOut.flagNoData = True |
|
|
1169 | 1169 | profileIndex = None |
|
|
1170 | 1170 | |
|
|
1171 | 1171 | if dataOut.flagDataAsBlock: |
|
|
1172 | 1172 | |
|
|
1173 | 1173 | dataOut.data = numpy.reshape(dataOut.data, shape_tuple) |
|
|
1174 | 1174 | dataOut.flagNoData = False |
|
|
1175 | 1175 | |
|
|
1176 | 1176 | profileIndex = int(dataOut.nProfiles*self.__nTxs) - 1 |
|
|
1177 | 1177 | |
|
|
1178 | 1178 | else: |
|
|
1179 | 1179 | |
|
|
1180 | 1180 | if self.__nTxs < 1: |
|
|
1181 | 1181 | |
|
|
1182 | 1182 | self.__appendProfile(dataOut, self.__nTxs) |
|
|
1183 | 1183 | new_data = self.__getBuffer() |
|
|
1184 | 1184 | |
|
|
1185 | 1185 | if new_data is not None: |
|
|
1186 | 1186 | dataOut.data = new_data |
|
|
1187 | 1187 | dataOut.flagNoData = False |
|
|
1188 | 1188 | |
|
|
1189 | 1189 | profileIndex = dataOut.profileIndex*nTxs |
|
|
1190 | 1190 | |
|
|
1191 | 1191 | else: |
|
|
1192 | 1192 | raise ValueError("nTxs should be greater than 0 and lower than 1, or use VoltageReader(..., getblock=True)") |
|
|
1193 | 1193 | |
|
|
1194 | 1194 | deltaHeight = dataOut.heightList[1] - dataOut.heightList[0] |
|
|
1195 | 1195 | |
|
|
1196 | 1196 | dataOut.heightList = numpy.arange(dataOut.nHeights/self.__nTxs) * deltaHeight + dataOut.heightList[0] |
|
|
1197 | 1197 | |
|
|
1198 | 1198 | dataOut.nProfiles = int(dataOut.nProfiles*self.__nTxs) |
|
|
1199 | 1199 | |
|
|
1200 | 1200 | dataOut.profileIndex = profileIndex |
|
|
1201 | 1201 | |
|
|
1202 | 1202 | dataOut.ippSeconds /= self.__nTxs |
|
|
1203 | 1203 | |
|
|
1204 | 1204 | return dataOut |
|
|
1205 | 1205 | |
|
|
1206 | 1206 | class SplitProfiles(Operation): |
|
|
1207 | 1207 | |
|
|
1208 | 1208 | def __init__(self, **kwargs): |
|
|
1209 | 1209 | |
|
|
1210 | 1210 | Operation.__init__(self, **kwargs) |
|
|
1211 | 1211 | |
|
|
1212 | 1212 | def run(self, dataOut, n): |
|
|
1213 | 1213 | |
|
|
1214 | 1214 | dataOut.flagNoData = True |
|
|
1215 | 1215 | profileIndex = None |
|
|
1216 | 1216 | |
|
|
1217 | 1217 | if dataOut.flagDataAsBlock: |
|
|
1218 | 1218 | |
|
|
1219 | 1219 | #nchannels, nprofiles, nsamples |
|
|
1220 | 1220 | shape = dataOut.data.shape |
|
|
1221 | 1221 | |
|
|
1222 | 1222 | if shape[2] % n != 0: |
|
|
1223 | 1223 | raise ValueError("Could not split the data, n=%d has to be multiple of %d" %(n, shape[2])) |
|
|
1224 | 1224 | |
|
|
1225 | 1225 | new_shape = shape[0], shape[1]*n, int(shape[2]/n) |
|
|
1226 | 1226 | |
|
|
1227 | 1227 | dataOut.data = numpy.reshape(dataOut.data, new_shape) |
|
|
1228 | 1228 | dataOut.flagNoData = False |
|
|
1229 | 1229 | |
|
|
1230 | 1230 | profileIndex = int(dataOut.nProfiles/n) - 1 |
|
|
1231 | 1231 | |
|
|
1232 | 1232 | else: |
|
|
1233 | 1233 | |
|
|
1234 | 1234 | raise ValueError("Could not split the data when is read Profile by Profile. Use VoltageReader(..., getblock=True)") |
|
|
1235 | 1235 | |
|
|
1236 | 1236 | deltaHeight = dataOut.heightList[1] - dataOut.heightList[0] |
|
|
1237 | 1237 | |
|
|
1238 | 1238 | dataOut.heightList = numpy.arange(dataOut.nHeights/n) * deltaHeight + dataOut.heightList[0] |
|
|
1239 | 1239 | |
|
|
1240 | 1240 | dataOut.nProfiles = int(dataOut.nProfiles*n) |
|
|
1241 | 1241 | |
|
|
1242 | 1242 | dataOut.profileIndex = profileIndex |
|
|
1243 | 1243 | |
|
|
1244 | 1244 | dataOut.ippSeconds /= n |
|
|
1245 | 1245 | |
|
|
1246 | 1246 | return dataOut |
|
|
1247 | 1247 | |
|
|
1248 | 1248 | class CombineProfiles(Operation): |
|
|
1249 | 1249 | def __init__(self, **kwargs): |
|
|
1250 | 1250 | |
|
|
1251 | 1251 | Operation.__init__(self, **kwargs) |
|
|
1252 | 1252 | |
|
|
1253 | 1253 | self.__remData = None |
|
|
1254 | 1254 | self.__profileIndex = 0 |
|
|
1255 | 1255 | |
|
|
1256 | 1256 | def run(self, dataOut, n): |
|
|
1257 | 1257 | |
|
|
1258 | 1258 | dataOut.flagNoData = True |
|
|
1259 | 1259 | profileIndex = None |
|
|
1260 | 1260 | |
|
|
1261 | 1261 | if dataOut.flagDataAsBlock: |
|
|
1262 | 1262 | |
|
|
1263 | 1263 | #nchannels, nprofiles, nsamples |
|
|
1264 | 1264 | shape = dataOut.data.shape |
|
|
1265 | 1265 | new_shape = shape[0], shape[1]/n, shape[2]*n |
|
|
1266 | 1266 | |
|
|
1267 | 1267 | if shape[1] % n != 0: |
|
|
1268 | 1268 | raise ValueError("Could not split the data, n=%d has to be multiple of %d" %(n, shape[1])) |
|
|
1269 | 1269 | |
|
|
1270 | 1270 | dataOut.data = numpy.reshape(dataOut.data, new_shape) |
|
|
1271 | 1271 | dataOut.flagNoData = False |
|
|
1272 | 1272 | |
|
|
1273 | 1273 | profileIndex = int(dataOut.nProfiles*n) - 1 |
|
|
1274 | 1274 | |
|
|
1275 | 1275 | else: |
|
|
1276 | 1276 | |
|
|
1277 | 1277 | #nchannels, nsamples |
|
|
1278 | 1278 | if self.__remData is None: |
|
|
1279 | 1279 | newData = dataOut.data |
|
|
1280 | 1280 | else: |
|
|
1281 | 1281 | newData = numpy.concatenate((self.__remData, dataOut.data), axis=1) |
|
|
1282 | 1282 | |
|
|
1283 | 1283 | self.__profileIndex += 1 |
|
|
1284 | 1284 | |
|
|
1285 | 1285 | if self.__profileIndex < n: |
|
|
1286 | 1286 | self.__remData = newData |
|
|
1287 | 1287 | #continue |
|
|
1288 | 1288 | return |
|
|
1289 | 1289 | |
|
|
1290 | 1290 | self.__profileIndex = 0 |
|
|
1291 | 1291 | self.__remData = None |
|
|
1292 | 1292 | |
|
|
1293 | 1293 | dataOut.data = newData |
|
|
1294 | 1294 | dataOut.flagNoData = False |
|
|
1295 | 1295 | |
|
|
1296 | 1296 | profileIndex = dataOut.profileIndex/n |
|
|
1297 | 1297 | |
|
|
1298 | 1298 | |
|
|
1299 | 1299 | deltaHeight = dataOut.heightList[1] - dataOut.heightList[0] |
|
|
1300 | 1300 | |
|
|
1301 | 1301 | dataOut.heightList = numpy.arange(dataOut.nHeights*n) * deltaHeight + dataOut.heightList[0] |
|
|
1302 | 1302 | |
|
|
1303 | 1303 | dataOut.nProfiles = int(dataOut.nProfiles/n) |
|
|
1304 | 1304 | |
|
|
1305 | 1305 | dataOut.profileIndex = profileIndex |
|
|
1306 | 1306 | |
|
|
1307 | 1307 | dataOut.ippSeconds *= n |
|
|
1308 | 1308 | |
|
|
1309 | 1309 | return dataOut |
|
|
1310 | 1310 | |
|
|
1311 | 1311 | class PulsePairVoltage(Operation): |
|
|
1312 | 1312 | ''' |
|
|
1313 | 1313 | Function PulsePair(Signal Power, Velocity) |
|
|
1314 | 1314 | The real component of Lag[0] provides Intensity Information |
|
|
1315 | 1315 | The imag component of Lag[1] Phase provides Velocity Information |
|
|
1316 | 1316 | |
|
|
1317 | 1317 | Configuration Parameters: |
|
|
1318 | 1318 | nPRF = Number of Several PRF |
|
|
1319 | 1319 | theta = Degree Azimuth angel Boundaries |
|
|
1320 | 1320 | |
|
|
1321 | 1321 | Input: |
|
|
1322 | 1322 | self.dataOut |
|
|
1323 | 1323 | lag[N] |
|
|
1324 | 1324 | Affected: |
|
|
1325 | 1325 | self.dataOut.spc |
|
|
1326 | 1326 | ''' |
|
|
1327 | 1327 | isConfig = False |
|
|
1328 | 1328 | __profIndex = 0 |
|
|
1329 | 1329 | __initime = None |
|
|
1330 | 1330 | __lastdatatime = None |
|
|
1331 | 1331 | __buffer = None |
|
|
1332 | 1332 | noise = None |
|
|
1333 | 1333 | __dataReady = False |
|
|
1334 | 1334 | n = None |
|
|
1335 | 1335 | __nch = 0 |
|
|
1336 | 1336 | __nHeis = 0 |
|
|
1337 | 1337 | removeDC = False |
|
|
1338 | 1338 | ipp = None |
|
|
1339 | 1339 | lambda_ = 0 |
|
|
1340 | 1340 | |
|
|
1341 | 1341 | def __init__(self,**kwargs): |
|
|
1342 | 1342 | Operation.__init__(self,**kwargs) |
|
|
1343 | 1343 | |
|
|
1344 | 1344 | def setup(self, dataOut, n = None, removeDC=False): |
|
|
1345 | 1345 | ''' |
|
|
1346 | 1346 | n= Numero de PRF's de entrada |
|
|
1347 | 1347 | ''' |
|
|
1348 | 1348 | self.__initime = None |
|
|
1349 | 1349 | self.__lastdatatime = 0 |
|
|
1350 | 1350 | self.__dataReady = False |
|
|
1351 | 1351 | self.__buffer = 0 |
|
|
1352 | 1352 | self.__profIndex = 0 |
|
|
1353 | 1353 | self.noise = None |
|
|
1354 | 1354 | self.__nch = dataOut.nChannels |
|
|
1355 | 1355 | self.__nHeis = dataOut.nHeights |
|
|
1356 | 1356 | self.removeDC = removeDC |
|
|
1357 | 1357 | self.lambda_ = 3.0e8/(9345.0e6) |
|
|
1358 | 1358 | self.ippSec = dataOut.ippSeconds |
|
|
1359 | 1359 | self.nCohInt = dataOut.nCohInt |
|
|
1360 | 1360 | |
|
|
1361 | 1361 | if n == None: |
|
|
1362 | 1362 | raise ValueError("n should be specified.") |
|
|
1363 | 1363 | |
|
|
1364 | 1364 | if n != None: |
|
|
1365 | 1365 | if n<2: |
|
|
1366 | 1366 | raise ValueError("n should be greater than 2") |
|
|
1367 | 1367 | |
|
|
1368 | 1368 | self.n = n |
|
|
1369 | 1369 | self.__nProf = n |
|
|
1370 | 1370 | |
|
|
1371 | 1371 | self.__buffer = numpy.zeros((dataOut.nChannels, |
|
|
1372 | 1372 | n, |
|
|
1373 | 1373 | dataOut.nHeights), |
|
|
1374 | 1374 | dtype='complex') |
|
|
1375 | 1375 | |
|
|
1376 | 1376 | def putData(self,data): |
|
|
1377 | 1377 | ''' |
|
|
1378 | 1378 | Add a profile to he __buffer and increase in one the __profiel Index |
|
|
1379 | 1379 | ''' |
|
|
1380 | 1380 | self.__buffer[:,self.__profIndex,:]= data |
|
|
1381 | 1381 | self.__profIndex += 1 |
|
|
1382 | 1382 | return |
|
|
1383 | 1383 | |
|
|
1384 | 1384 | def pushData(self,dataOut): |
|
|
1385 | 1385 | ''' |
|
|
1386 | 1386 | Return the PULSEPAIR and the profiles used in the operation |
|
|
1387 | 1387 | Affected : self.__profileIndex |
|
|
1388 | 1388 | ''' |
|
|
1389 | 1389 | #----------------- Remove DC----------------------------------- |
|
|
1390 | 1390 | if self.removeDC==True: |
|
|
1391 | 1391 | mean = numpy.mean(self.__buffer,1) |
|
|
1392 | 1392 | tmp = mean.reshape(self.__nch,1,self.__nHeis) |
|
|
1393 | 1393 | dc= numpy.tile(tmp,[1,self.__nProf,1]) |
|
|
1394 | 1394 | self.__buffer = self.__buffer - dc |
|
|
1395 | 1395 | #------------------Calculo de Potencia ------------------------ |
|
|
1396 | 1396 | pair0 = self.__buffer*numpy.conj(self.__buffer) |
|
|
1397 | 1397 | pair0 = pair0.real |
|
|
1398 | 1398 | lag_0 = numpy.sum(pair0,1) |
|
|
1399 | 1399 | #------------------Calculo de Ruido x canal-------------------- |
|
|
1400 | 1400 | self.noise = numpy.zeros(self.__nch) |
|
|
1401 | 1401 | for i in range(self.__nch): |
|
|
1402 | 1402 | daux = numpy.sort(pair0[i,:,:],axis= None) |
|
|
1403 | 1403 | self.noise[i]=hildebrand_sekhon( daux ,self.nCohInt) |
|
|
1404 | 1404 | |
|
|
1405 | 1405 | self.noise = self.noise.reshape(self.__nch,1) |
|
|
1406 | 1406 | self.noise = numpy.tile(self.noise,[1,self.__nHeis]) |
|
|
1407 | 1407 | noise_buffer = self.noise.reshape(self.__nch,1,self.__nHeis) |
|
|
1408 | 1408 | noise_buffer = numpy.tile(noise_buffer,[1,self.__nProf,1]) |
|
|
1409 | 1409 | #------------------ Potencia recibida= P , Potencia senal = S , Ruido= N-- |
|
|
1410 | 1410 | #------------------ P= S+N ,P=lag_0/N --------------------------------- |
|
|
1411 | 1411 | #-------------------- Power -------------------------------------------------- |
|
|
1412 | 1412 | data_power = lag_0/(self.n*self.nCohInt) |
|
|
1413 | 1413 | #------------------ Senal --------------------------------------------------- |
|
|
1414 | 1414 | data_intensity = pair0 - noise_buffer |
|
|
1415 | 1415 | data_intensity = numpy.sum(data_intensity,axis=1)*(self.n*self.nCohInt)#*self.nCohInt) |
|
|
1416 | 1416 | #data_intensity = (lag_0-self.noise*self.n)*(self.n*self.nCohInt) |
|
|
1417 | 1417 | for i in range(self.__nch): |
|
|
1418 | 1418 | for j in range(self.__nHeis): |
|
|
1419 | 1419 | if data_intensity[i][j] < 0: |
|
|
1420 | 1420 | data_intensity[i][j] = numpy.min(numpy.absolute(data_intensity[i][j])) |
|
|
1421 | 1421 | |
|
|
1422 | 1422 | #----------------- Calculo de Frecuencia y Velocidad doppler-------- |
|
|
1423 | 1423 | pair1 = self.__buffer[:,:-1,:]*numpy.conjugate(self.__buffer[:,1:,:]) |
|
|
1424 | 1424 | lag_1 = numpy.sum(pair1,1) |
|
|
1425 | 1425 | data_freq = (-1/(2.0*math.pi*self.ippSec*self.nCohInt))*numpy.angle(lag_1) |
|
|
1426 | 1426 | data_velocity = (self.lambda_/2.0)*data_freq |
|
|
1427 | 1427 | |
|
|
1428 | 1428 | #---------------- Potencia promedio estimada de la Senal----------- |
|
|
1429 | 1429 | lag_0 = lag_0/self.n |
|
|
1430 | 1430 | S = lag_0-self.noise |
|
|
1431 | 1431 | |
|
|
1432 | 1432 | #---------------- Frecuencia Doppler promedio --------------------- |
|
|
1433 | 1433 | lag_1 = lag_1/(self.n-1) |
|
|
1434 | 1434 | R1 = numpy.abs(lag_1) |
|
|
1435 | 1435 | |
|
|
1436 | 1436 | #---------------- Calculo del SNR---------------------------------- |
|
|
1437 | 1437 | data_snrPP = S/self.noise |
|
|
1438 | 1438 | for i in range(self.__nch): |
|
|
1439 | 1439 | for j in range(self.__nHeis): |
|
|
1440 | 1440 | if data_snrPP[i][j] < 1.e-20: |
|
|
1441 | 1441 | data_snrPP[i][j] = 1.e-20 |
|
|
1442 | 1442 | |
|
|
1443 | 1443 | #----------------- Calculo del ancho espectral ---------------------- |
|
|
1444 | 1444 | L = S/R1 |
|
|
1445 | 1445 | L = numpy.where(L<0,1,L) |
|
|
1446 | 1446 | L = numpy.log(L) |
|
|
1447 | 1447 | tmp = numpy.sqrt(numpy.absolute(L)) |
|
|
1448 | 1448 | data_specwidth = (self.lambda_/(2*math.sqrt(2)*math.pi*self.ippSec*self.nCohInt))*tmp*numpy.sign(L) |
|
|
1449 | 1449 | n = self.__profIndex |
|
|
1450 | 1450 | |
|
|
1451 | 1451 | self.__buffer = numpy.zeros((self.__nch, self.__nProf,self.__nHeis), dtype='complex') |
|
|
1452 | 1452 | self.__profIndex = 0 |
|
|
1453 | 1453 | return data_power,data_intensity,data_velocity,data_snrPP,data_specwidth,n |
|
|
1454 | 1454 | |
|
|
1455 | 1455 | |
|
|
1456 | 1456 | def pulsePairbyProfiles(self,dataOut): |
|
|
1457 | 1457 | |
|
|
1458 | 1458 | self.__dataReady = False |
|
|
1459 | 1459 | data_power = None |
|
|
1460 | 1460 | data_intensity = None |
|
|
1461 | 1461 | data_velocity = None |
|
|
1462 | 1462 | data_specwidth = None |
|
|
1463 | 1463 | data_snrPP = None |
|
|
1464 | 1464 | self.putData(data=dataOut.data) |
|
|
1465 | 1465 | if self.__profIndex == self.n: |
|
|
1466 | 1466 | data_power,data_intensity, data_velocity,data_snrPP,data_specwidth, n = self.pushData(dataOut=dataOut) |
|
|
1467 | 1467 | self.__dataReady = True |
|
|
1468 | 1468 | |
|
|
1469 | 1469 | return data_power, data_intensity, data_velocity, data_snrPP, data_specwidth |
|
|
1470 | 1470 | |
|
|
1471 | 1471 | |
|
|
1472 | 1472 | def pulsePairOp(self, dataOut, datatime= None): |
|
|
1473 | 1473 | |
|
|
1474 | 1474 | if self.__initime == None: |
|
|
1475 | 1475 | self.__initime = datatime |
|
|
1476 | 1476 | data_power, data_intensity, data_velocity, data_snrPP, data_specwidth = self.pulsePairbyProfiles(dataOut) |
|
|
1477 | 1477 | self.__lastdatatime = datatime |
|
|
1478 | 1478 | |
|
|
1479 | 1479 | if data_power is None: |
|
|
1480 | 1480 | return None, None, None,None,None,None |
|
|
1481 | 1481 | |
|
|
1482 | 1482 | avgdatatime = self.__initime |
|
|
1483 | 1483 | deltatime = datatime - self.__lastdatatime |
|
|
1484 | 1484 | self.__initime = datatime |
|
|
1485 | 1485 | |
|
|
1486 | 1486 | return data_power, data_intensity, data_velocity, data_snrPP, data_specwidth, avgdatatime |
|
|
1487 | 1487 | |
|
|
1488 | 1488 | def run(self, dataOut,n = None,removeDC= False, overlapping= False,**kwargs): |
|
|
1489 | 1489 | |
|
|
1490 | 1490 | if not self.isConfig: |
|
|
1491 | 1491 | self.setup(dataOut = dataOut, n = n , removeDC=removeDC , **kwargs) |
|
|
1492 | 1492 | self.isConfig = True |
|
|
1493 | 1493 | data_power, data_intensity, data_velocity,data_snrPP,data_specwidth, avgdatatime = self.pulsePairOp(dataOut, dataOut.utctime) |
|
|
1494 | 1494 | dataOut.flagNoData = True |
|
|
1495 | 1495 | |
|
|
1496 | 1496 | if self.__dataReady: |
|
|
1497 | 1497 | dataOut.nCohInt *= self.n |
|
|
1498 | 1498 | dataOut.dataPP_POW = data_intensity # S |
|
|
1499 | 1499 | dataOut.dataPP_POWER = data_power # P |
|
|
1500 | 1500 | dataOut.dataPP_DOP = data_velocity |
|
|
1501 | 1501 | dataOut.dataPP_SNR = data_snrPP |
|
|
1502 | 1502 | dataOut.dataPP_WIDTH = data_specwidth |
|
|
1503 | 1503 | dataOut.PRFbyAngle = self.n #numero de PRF*cada angulo rotado que equivale a un tiempo. |
|
|
1504 | 1504 | dataOut.utctime = avgdatatime |
|
|
1505 | 1505 | dataOut.flagNoData = False |
|
|
1506 | 1506 | return dataOut |
|
|
1507 | 1507 | |
|
|
1508 | 1508 | |
|
|
1509 | 1509 | |
|
|
1510 | 1510 | # import collections |
|
|
1511 | 1511 | # from scipy.stats import mode |
|
|
1512 | 1512 | # |
|
|
1513 | 1513 | # class Synchronize(Operation): |
|
|
1514 | 1514 | # |
|
|
1515 | 1515 | # isConfig = False |
|
|
1516 | 1516 | # __profIndex = 0 |
|
|
1517 | 1517 | # |
|
|
1518 | 1518 | # def __init__(self, **kwargs): |
|
|
1519 | 1519 | # |
|
|
1520 | 1520 | # Operation.__init__(self, **kwargs) |
|
|
1521 | 1521 | # # self.isConfig = False |
|
|
1522 | 1522 | # self.__powBuffer = None |
|
|
1523 | 1523 | # self.__startIndex = 0 |
|
|
1524 | 1524 | # self.__pulseFound = False |
|
|
1525 | 1525 | # |
|
|
1526 | 1526 | # def __findTxPulse(self, dataOut, channel=0, pulse_with = None): |
|
|
1527 | 1527 | # |
|
|
1528 | 1528 | # #Read data |
|
|
1529 | 1529 | # |
|
|
1530 | 1530 | # powerdB = dataOut.getPower(channel = channel) |
|
|
1531 | 1531 | # noisedB = dataOut.getNoise(channel = channel)[0] |
|
|
1532 | 1532 | # |
|
|
1533 | 1533 | # self.__powBuffer.extend(powerdB.flatten()) |
|
|
1534 | 1534 | # |
|
|
1535 | 1535 | # dataArray = numpy.array(self.__powBuffer) |
|
|
1536 | 1536 | # |
|
|
1537 | 1537 | # filteredPower = numpy.correlate(dataArray, dataArray[0:self.__nSamples], "same") |
|
|
1538 | 1538 | # |
|
|
1539 | 1539 | # maxValue = numpy.nanmax(filteredPower) |
|
|
1540 | 1540 | # |
|
|
1541 | 1541 | # if maxValue < noisedB + 10: |
|
|
1542 | 1542 | # #No se encuentra ningun pulso de transmision |
|
|
1543 | 1543 | # return None |
|
|
1544 | 1544 | # |
|
|
1545 | 1545 | # maxValuesIndex = numpy.where(filteredPower > maxValue - 0.1*abs(maxValue))[0] |
|
|
1546 | 1546 | # |
|
|
1547 | 1547 | # if len(maxValuesIndex) < 2: |
|
|
1548 | 1548 | # #Solo se encontro un solo pulso de transmision de un baudio, esperando por el siguiente TX |
|
|
1549 | 1549 | # return None |
|
|
1550 | 1550 | # |
|
|
1551 | 1551 | # phasedMaxValuesIndex = maxValuesIndex - self.__nSamples |
|
|
1552 | 1552 | # |
|
|
1553 | 1553 | # #Seleccionar solo valores con un espaciamiento de nSamples |
|
|
1554 | 1554 | # pulseIndex = numpy.intersect1d(maxValuesIndex, phasedMaxValuesIndex) |
|
|
1555 | 1555 | # |
|
|
1556 | 1556 | # if len(pulseIndex) < 2: |
|
|
1557 | 1557 | # #Solo se encontro un pulso de transmision con ancho mayor a 1 |
|
|
1558 | 1558 | # return None |
|
|
1559 | 1559 | # |
|
|
1560 | 1560 | # spacing = pulseIndex[1:] - pulseIndex[:-1] |
|
|
1561 | 1561 | # |
|
|
1562 | 1562 | # #remover senales que se distancien menos de 10 unidades o muestras |
|
|
1563 | 1563 | # #(No deberian existir IPP menor a 10 unidades) |
|
|
1564 | 1564 | # |
|
|
1565 | 1565 | # realIndex = numpy.where(spacing > 10 )[0] |
|
|
1566 | 1566 | # |
|
|
1567 | 1567 | # if len(realIndex) < 2: |
|
|
1568 | 1568 | # #Solo se encontro un pulso de transmision con ancho mayor a 1 |
|
|
1569 | 1569 | # return None |
|
|
1570 | 1570 | # |
|
|
1571 | 1571 | # #Eliminar pulsos anchos (deja solo la diferencia entre IPPs) |
|
|
1572 | 1572 | # realPulseIndex = pulseIndex[realIndex] |
|
|
1573 | 1573 | # |
|
|
1574 | 1574 | # period = mode(realPulseIndex[1:] - realPulseIndex[:-1])[0][0] |
|
|
1575 | 1575 | # |
|
|
1576 | 1576 | # print "IPP = %d samples" %period |
|
|
1577 | 1577 | # |
|
|
1578 | 1578 | # self.__newNSamples = dataOut.nHeights #int(period) |
|
|
1579 | 1579 | # self.__startIndex = int(realPulseIndex[0]) |
|
|
1580 | 1580 | # |
|
|
1581 | 1581 | # return 1 |
|
|
1582 | 1582 | # |
|
|
1583 | 1583 | # |
|
|
1584 | 1584 | # def setup(self, nSamples, nChannels, buffer_size = 4): |
|
|
1585 | 1585 | # |
|
|
1586 | 1586 | # self.__powBuffer = collections.deque(numpy.zeros( buffer_size*nSamples,dtype=numpy.float), |
|
|
1587 | 1587 | # maxlen = buffer_size*nSamples) |
|
|
1588 | 1588 | # |
|
|
1589 | 1589 | # bufferList = [] |
|
|
1590 | 1590 | # |
|
|
1591 | 1591 | # for i in range(nChannels): |
|
|
1592 | 1592 | # bufferByChannel = collections.deque(numpy.zeros( buffer_size*nSamples, dtype=numpy.complex) + numpy.NAN, |
|
|
1593 | 1593 | # maxlen = buffer_size*nSamples) |
|
|
1594 | 1594 | # |
|
|
1595 | 1595 | # bufferList.append(bufferByChannel) |
|
|
1596 | 1596 | # |
|
|
1597 | 1597 | # self.__nSamples = nSamples |
|
|
1598 | 1598 | # self.__nChannels = nChannels |
|
|
1599 | 1599 | # self.__bufferList = bufferList |
|
|
1600 | 1600 | # |
|
|
1601 | 1601 | # def run(self, dataOut, channel = 0): |
|
|
1602 | 1602 | # |
|
|
1603 | 1603 | # if not self.isConfig: |
|
|
1604 | 1604 | # nSamples = dataOut.nHeights |
|
|
1605 | 1605 | # nChannels = dataOut.nChannels |
|
|
1606 | 1606 | # self.setup(nSamples, nChannels) |
|
|
1607 | 1607 | # self.isConfig = True |
|
|
1608 | 1608 | # |
|
|
1609 | 1609 | # #Append new data to internal buffer |
|
|
1610 | 1610 | # for thisChannel in range(self.__nChannels): |
|
|
1611 | 1611 | # bufferByChannel = self.__bufferList[thisChannel] |
|
|
1612 | 1612 | # bufferByChannel.extend(dataOut.data[thisChannel]) |
|
|
1613 | 1613 | # |
|
|
1614 | 1614 | # if self.__pulseFound: |
|
|
1615 | 1615 | # self.__startIndex -= self.__nSamples |
|
|
1616 | 1616 | # |
|
|
1617 | 1617 | # #Finding Tx Pulse |
|
|
1618 | 1618 | # if not self.__pulseFound: |
|
|
1619 | 1619 | # indexFound = self.__findTxPulse(dataOut, channel) |
|
|
1620 | 1620 | # |
|
|
1621 | 1621 | # if indexFound == None: |
|
|
1622 | 1622 | # dataOut.flagNoData = True |
|
|
1623 | 1623 | # return |
|
|
1624 | 1624 | # |
|
|
1625 | 1625 | # self.__arrayBuffer = numpy.zeros((self.__nChannels, self.__newNSamples), dtype = numpy.complex) |
|
|
1626 | 1626 | # self.__pulseFound = True |
|
|
1627 | 1627 | # self.__startIndex = indexFound |
|
|
1628 | 1628 | # |
|
|
1629 | 1629 | # #If pulse was found ... |
|
|
1630 | 1630 | # for thisChannel in range(self.__nChannels): |
|
|
1631 | 1631 | # bufferByChannel = self.__bufferList[thisChannel] |
|
|
1632 | 1632 | # #print self.__startIndex |
|
|
1633 | 1633 | # x = numpy.array(bufferByChannel) |
|
|
1634 | 1634 | # self.__arrayBuffer[thisChannel] = x[self.__startIndex:self.__startIndex+self.__newNSamples] |
|
|
1635 | 1635 | # |
|
|
1636 | 1636 | # deltaHeight = dataOut.heightList[1] - dataOut.heightList[0] |
|
|
1637 | 1637 | # dataOut.heightList = numpy.arange(self.__newNSamples)*deltaHeight |
|
|
1638 | 1638 | # # dataOut.ippSeconds = (self.__newNSamples / deltaHeight)/1e6 |
|
|
1639 | 1639 | # |
|
|
1640 | 1640 | # dataOut.data = self.__arrayBuffer |
|
|
1641 | 1641 | # |
|
|
1642 | 1642 | # self.__startIndex += self.__newNSamples |
|
|
1643 | 1643 | # |
|
|
1644 | 1644 | # return |
|
|
1645 | 1645 | class SSheightProfiles(Operation): |
|
|
1646 | 1646 | |
|
|
1647 | 1647 | step = None |
|
|
1648 | 1648 | nsamples = None |
|
|
1649 | 1649 | bufferShape = None |
|
|
1650 | 1650 | profileShape = None |
|
|
1651 | 1651 | sshProfiles = None |
|
|
1652 | 1652 | profileIndex = None |
|
|
1653 | 1653 | |
|
|
1654 | 1654 | def __init__(self, **kwargs): |
|
|
1655 | 1655 | |
|
|
1656 | 1656 | Operation.__init__(self, **kwargs) |
|
|
1657 | 1657 | self.isConfig = False |
|
|
1658 | 1658 | |
|
|
1659 | 1659 | def setup(self,dataOut ,step = None , nsamples = None): |
|
|
1660 | 1660 | |
|
|
1661 | 1661 | if step == None and nsamples == None: |
|
|
1662 | 1662 | raise ValueError("step or nheights should be specified ...") |
|
|
1663 | 1663 | |
|
|
1664 | 1664 | self.step = step |
|
|
1665 | 1665 | self.nsamples = nsamples |
|
|
1666 | 1666 | self.__nChannels = dataOut.nChannels |
|
|
1667 | 1667 | self.__nProfiles = dataOut.nProfiles |
|
|
1668 | 1668 | self.__nHeis = dataOut.nHeights |
|
|
1669 | 1669 | shape = dataOut.data.shape #nchannels, nprofiles, nsamples |
|
|
1670 | 1670 | |
|
|
1671 | 1671 | residue = (shape[1] - self.nsamples) % self.step |
|
|
1672 | 1672 | if residue != 0: |
|
|
1673 | 1673 | print("The residue is %d, step=%d should be multiple of %d to avoid loss of %d samples"%(residue,step,shape[1] - self.nsamples,residue)) |
|
|
1674 | 1674 | |
|
|
1675 | 1675 | deltaHeight = dataOut.heightList[1] - dataOut.heightList[0] |
|
|
1676 | 1676 | numberProfile = self.nsamples |
|
|
1677 | 1677 | numberSamples = (shape[1] - self.nsamples)/self.step |
|
|
1678 | 1678 | |
|
|
1679 | 1679 | self.bufferShape = int(shape[0]), int(numberSamples), int(numberProfile) # nchannels, nsamples , nprofiles |
|
|
1680 | 1680 | self.profileShape = int(shape[0]), int(numberProfile), int(numberSamples) # nchannels, nprofiles, nsamples |
|
|
1681 | 1681 | |
|
|
1682 | 1682 | self.buffer = numpy.zeros(self.bufferShape , dtype=numpy.complex) |
|
|
1683 | 1683 | self.sshProfiles = numpy.zeros(self.profileShape, dtype=numpy.complex) |
|
|
1684 | 1684 | |
|
|
1685 | 1685 | def run(self, dataOut, step, nsamples, code = None, repeat = None): |
|
|
1686 | 1686 | dataOut.flagNoData = True |
|
|
1687 | 1687 | |
|
|
1688 | 1688 | profileIndex = None |
|
|
1689 | 1689 | #print("nProfiles, nHeights ",dataOut.nProfiles, dataOut.nHeights) |
|
|
1690 | 1690 | #print(dataOut.getFreqRange(1)/1000.) |
|
|
1691 | 1691 | #exit(1) |
|
|
1692 | 1692 | if dataOut.flagDataAsBlock: |
|
|
1693 | 1693 | dataOut.data = numpy.average(dataOut.data,axis=1) |
|
|
1694 | 1694 | #print("jee") |
|
|
1695 | 1695 | dataOut.flagDataAsBlock = False |
|
|
1696 | 1696 | if not self.isConfig: |
|
|
1697 | 1697 | self.setup(dataOut, step=step , nsamples=nsamples) |
|
|
1698 | 1698 | #print("Setup done") |
|
|
1699 | 1699 | self.isConfig = True |
|
|
1700 | 1700 | |
|
|
1701 | 1701 | |
|
|
1702 | 1702 | if code is not None: |
|
|
1703 | 1703 | code = numpy.array(code) |
|
|
1704 | 1704 | code_block = code |
|
|
1705 | 1705 | |
|
|
1706 | 1706 | if repeat is not None: |
|
|
1707 | 1707 | code_block = numpy.repeat(code_block, repeats=repeat, axis=1) |
|
|
1708 | 1708 | #print(code_block.shape) |
|
|
1709 | 1709 | for i in range(self.buffer.shape[1]): |
|
|
1710 | 1710 | |
|
|
1711 | 1711 | if code is not None: |
|
|
1712 | 1712 | self.buffer[:,i] = dataOut.data[:,i*self.step:i*self.step + self.nsamples]*code_block |
|
|
1713 | 1713 | |
|
|
1714 | 1714 | else: |
|
|
1715 | 1715 | |
|
|
1716 | 1716 | self.buffer[:,i] = dataOut.data[:,i*self.step:i*self.step + self.nsamples]#*code[dataOut.profileIndex,:] |
|
|
1717 | 1717 | |
|
|
1718 | 1718 | #self.buffer[:,j,self.__nHeis-j*self.step - self.nheights:self.__nHeis-j*self.step] = numpy.flip(dataOut.data[:,j*self.step:j*self.step + self.nheights]) |
|
|
1719 | 1719 | |
|
|
1720 | 1720 | for j in range(self.buffer.shape[0]): |
|
|
1721 | 1721 | self.sshProfiles[j] = numpy.transpose(self.buffer[j]) |
|
|
1722 | 1722 | |
|
|
1723 | 1723 | profileIndex = self.nsamples |
|
|
1724 | 1724 | deltaHeight = dataOut.heightList[1] - dataOut.heightList[0] |
|
|
1725 | 1725 | ippSeconds = (deltaHeight*1.0e-6)/(0.15) |
|
|
1726 | 1726 | #print("ippSeconds, dH: ",ippSeconds,deltaHeight) |
|
|
1727 | 1727 | try: |
|
|
1728 | 1728 | if dataOut.concat_m is not None: |
|
|
1729 | 1729 | ippSeconds= ippSeconds/float(dataOut.concat_m) |
|
|
1730 | 1730 | #print "Profile concat %d"%dataOut.concat_m |
|
|
1731 | 1731 | except: |
|
|
1732 | 1732 | pass |
|
|
1733 | 1733 | |
|
|
1734 | 1734 | dataOut.data = self.sshProfiles |
|
|
1735 | 1735 | dataOut.flagNoData = False |
|
|
1736 | 1736 | dataOut.heightList = numpy.arange(self.buffer.shape[1]) *self.step*deltaHeight + dataOut.heightList[0] |
|
|
1737 | 1737 | dataOut.nProfiles = int(dataOut.nProfiles*self.nsamples) |
|
|
1738 | 1738 | |
|
|
1739 | 1739 | dataOut.profileIndex = profileIndex |
|
|
1740 | 1740 | dataOut.flagDataAsBlock = True |
|
|
1741 | 1741 | dataOut.ippSeconds = ippSeconds |
|
|
1742 | 1742 | dataOut.step = self.step |
|
|
1743 | 1743 | #print(numpy.shape(dataOut.data)) |
|
|
1744 | 1744 | #exit(1) |
|
|
1745 | 1745 | #print("new data shape and time:", dataOut.data.shape, dataOut.utctime) |
|
|
1746 | 1746 | |
|
|
1747 | 1747 | return dataOut |
|
|
1748 | 1748 | ################################################################################3############################3 |
|
|
1749 | 1749 | ################################################################################3############################3 |
|
|
1750 | 1750 | ################################################################################3############################3 |
|
|
1751 | 1751 | ################################################################################3############################3 |
|
|
1752 | 1752 | |
|
|
1753 | 1753 | class SSheightProfiles2(Operation): |
|
|
1754 | 1754 | ''' |
|
|
1755 | 1755 | Procesa por perfiles y por bloques |
|
|
1756 | 1756 | ''' |
|
|
1757 | 1757 | |
|
|
1758 | 1758 | |
|
|
1759 | 1759 | bufferShape = None |
|
|
1760 | 1760 | profileShape = None |
|
|
1761 | 1761 | sshProfiles = None |
|
|
1762 | 1762 | profileIndex = None |
|
|
1763 | 1763 | #nsamples = None |
|
|
1764 | 1764 | #step = None |
|
|
1765 | 1765 | #deltaHeight = None |
|
|
1766 | 1766 | #init_range = None |
|
|
1767 | 1767 | __slots__ = ('step', 'nsamples', 'deltaHeight', 'init_range', 'isConfig', '__nChannels', |
|
|
1768 | 1768 | '__nProfiles', '__nHeis', 'deltaHeight', 'new_nHeights') |
|
|
1769 | 1769 | |
|
|
1770 | 1770 | def __init__(self, **kwargs): |
|
|
1771 | 1771 | |
|
|
1772 | 1772 | Operation.__init__(self, **kwargs) |
|
|
1773 | 1773 | self.isConfig = False |
|
|
1774 | 1774 | |
|
|
1775 | 1775 | def setup(self,dataOut ,step = None , nsamples = None): |
|
|
1776 | 1776 | |
|
|
1777 | 1777 | if step == None and nsamples == None: |
|
|
1778 | 1778 | raise ValueError("step or nheights should be specified ...") |
|
|
1779 | 1779 | |
|
|
1780 | 1780 | self.step = step |
|
|
1781 | 1781 | self.nsamples = nsamples |
|
|
1782 | 1782 | self.__nChannels = int(dataOut.nChannels) |
|
|
1783 | 1783 | self.__nProfiles = int(dataOut.nProfiles) |
|
|
1784 | 1784 | self.__nHeis = int(dataOut.nHeights) |
|
|
1785 | 1785 | |
|
|
1786 | 1786 | residue = (self.__nHeis - self.nsamples) % self.step |
|
|
1787 | 1787 | if residue != 0: |
|
|
1788 | 1788 | print("The residue is %d, step=%d should be multiple of %d to avoid loss of %d samples"%(residue,step,shape[1] - self.nsamples,residue)) |
|
|
1789 | 1789 | |
|
|
1790 | 1790 | self.deltaHeight = dataOut.heightList[1] - dataOut.heightList[0] |
|
|
1791 | 1791 | self.init_range = dataOut.heightList[0] |
|
|
1792 | 1792 | #numberProfile = self.nsamples |
|
|
1793 | 1793 | numberSamples = (self.__nHeis - self.nsamples)/self.step |
|
|
1794 | 1794 | |
|
|
1795 | 1795 | self.new_nHeights = numberSamples |
|
|
1796 | 1796 | |
|
|
1797 | 1797 | self.bufferShape = int(self.__nChannels), int(numberSamples), int(self.nsamples) # nchannels, nsamples , nprofiles |
|
|
1798 | 1798 | self.profileShape = int(self.__nChannels), int(self.nsamples), int(numberSamples) # nchannels, nprofiles, nsamples |
|
|
1799 | 1799 | |
|
|
1800 | 1800 | self.buffer = numpy.zeros(self.bufferShape , dtype=numpy.complex) |
|
|
1801 | 1801 | self.sshProfiles = numpy.zeros(self.profileShape, dtype=numpy.complex) |
|
|
1802 | 1802 | |
|
|
1803 | 1803 | def getNewProfiles(self, data, code=None, repeat=None): |
|
|
1804 | 1804 | |
|
|
1805 | 1805 | if code is not None: |
|
|
1806 | 1806 | code = numpy.array(code) |
|
|
1807 | 1807 | code_block = code |
|
|
1808 | 1808 | |
|
|
1809 | 1809 | if repeat is not None: |
|
|
1810 | 1810 | code_block = numpy.repeat(code_block, repeats=repeat, axis=1) |
|
|
1811 | 1811 | if data.ndim == 2: |
|
|
1812 | 1812 | data = data.reshape(1,1,self.__nHeis ) |
|
|
1813 | 1813 | #print("buff, data, :",self.buffer.shape, data.shape,self.sshProfiles.shape) |
|
|
1814 | 1814 | for i in range(int(self.new_nHeights)): #nuevas alturas |
|
|
1815 | 1815 | if code is not None: |
|
|
1816 | 1816 | self.buffer[:,i,:] = data[:,:,i*self.step:i*self.step + self.nsamples]*code_block |
|
|
1817 | 1817 | else: |
|
|
1818 | 1818 | self.buffer[:,i,:] = data[:,:,i*self.step:i*self.step + self.nsamples]#*code[dataOut.profileIndex,:] |
|
|
1819 | 1819 | |
|
|
1820 | 1820 | for j in range(self.__nChannels): #en los cananles |
|
|
1821 | 1821 | self.sshProfiles[j,:,:] = numpy.transpose(self.buffer[j,:,:]) |
|
|
1822 | 1822 | #print("new profs Done") |
|
|
1823 | 1823 | |
|
|
1824 | 1824 | |
|
|
1825 | 1825 | |
|
|
1826 | 1826 | def run(self, dataOut, step, nsamples, code = None, repeat = None): |
|
|
1827 | 1827 | |
|
|
1828 | 1828 | if dataOut.flagNoData == True: |
|
|
1829 | 1829 | return dataOut |
|
|
1830 | 1830 | dataOut.flagNoData = True |
|
|
1831 | 1831 | #print("init data shape:", dataOut.data.shape) |
|
|
1832 | 1832 | #print("ch: {} prof: {} hs: {}".format(int(dataOut.nChannels), |
|
|
1833 | 1833 | # int(dataOut.nProfiles),int(dataOut.nHeights))) |
|
|
1834 | 1834 | |
|
|
1835 | 1835 | profileIndex = None |
|
|
1836 | 1836 | # if not dataOut.flagDataAsBlock: |
|
|
1837 | 1837 | # dataOut.nProfiles = 1 |
|
|
1838 | 1838 | |
|
|
1839 | 1839 | if not self.isConfig: |
|
|
1840 | 1840 | self.setup(dataOut, step=step , nsamples=nsamples) |
|
|
1841 | 1841 | #print("Setup done") |
|
|
1842 | 1842 | self.isConfig = True |
|
|
1843 | 1843 | |
|
|
1844 | 1844 | dataBlock = None |
|
|
1845 | 1845 | |
|
|
1846 | 1846 | nprof = 1 |
|
|
1847 | 1847 | if dataOut.flagDataAsBlock: |
|
|
1848 | 1848 | nprof = int(dataOut.nProfiles) |
|
|
1849 | 1849 | |
|
|
1850 | 1850 | #print("dataOut nProfiles:", dataOut.nProfiles) |
|
|
1851 | 1851 | for profile in range(nprof): |
|
|
1852 | 1852 | if dataOut.flagDataAsBlock: |
|
|
1853 | 1853 | #print("read blocks") |
|
|
1854 | 1854 | self.getNewProfiles(dataOut.data[:,profile,:], code=code, repeat=repeat) |
|
|
1855 | 1855 | else: |
|
|
1856 | 1856 | #print("read profiles") |
|
|
1857 | 1857 | self.getNewProfiles(dataOut.data, code=code, repeat=repeat) #only one channe |
|
|
1858 | 1858 | if profile == 0: |
|
|
1859 | 1859 | dataBlock = self.sshProfiles.copy() |
|
|
1860 | 1860 | else: #by blocks |
|
|
1861 | 1861 | dataBlock = numpy.concatenate((dataBlock,self.sshProfiles), axis=1) #profile axis |
|
|
1862 | 1862 | #print("by blocks: ",dataBlock.shape, self.sshProfiles.shape) |
|
|
1863 | 1863 | |
|
|
1864 | 1864 | profileIndex = self.nsamples |
|
|
1865 | 1865 | #deltaHeight = dataOut.heightList[1] - dataOut.heightList[0] |
|
|
1866 | 1866 | ippSeconds = (self.deltaHeight*1.0e-6)/(0.15) |
|
|
1867 | 1867 | |
|
|
1868 | 1868 | |
|
|
1869 | 1869 | dataOut.data = dataBlock |
|
|
1870 | 1870 | #print("show me: ",self.step,self.deltaHeight, dataOut.heightList, self.new_nHeights) |
|
|
1871 | 1871 | dataOut.heightList = numpy.arange(int(self.new_nHeights)) *self.step*self.deltaHeight + self.init_range |
|
|
1872 | 1872 | |
|
|
1873 | 1873 | dataOut.ippSeconds = ippSeconds |
|
|
1874 | 1874 | dataOut.step = self.step |
|
|
1875 | 1875 | dataOut.flagNoData = False |
|
|
1876 | 1876 | if dataOut.flagDataAsBlock: |
|
|
1877 | 1877 | dataOut.nProfiles = int(dataOut.nProfiles*self.nsamples) |
|
|
1878 | 1878 | |
|
|
1879 | 1879 | else: |
|
|
1880 | 1880 | dataOut.nProfiles = int(self.nsamples) |
|
|
1881 | 1881 | dataOut.profileIndex = dataOut.nProfiles |
|
|
1882 | 1882 | dataOut.flagDataAsBlock = True |
|
|
1883 | 1883 | |
|
|
1884 | 1884 | dataBlock = None |
|
|
1885 | 1885 | |
|
|
1886 | 1886 | #print("new data shape:", dataOut.data.shape, dataOut.utctime) |
|
|
1887 | 1887 | |
|
|
1888 | 1888 | return dataOut |
|
|
1889 | 1889 | |
|
|
1890 | 1890 | |
|
|
1891 | 1891 | |
|
|
1892 | 1892 | |
|
|
1893 | #import skimage.color | |
|
|
1894 | #import skimage.io | |
|
|
1895 | #import matplotlib.pyplot as plt | |
|
|
1896 | 1893 | |
|
|
1897 | 1894 | class removeProfileByFaradayHS(Operation): |
|
|
1898 | 1895 | ''' |
|
|
1899 | 1896 | |
|
|
1900 | 1897 | ''' |
|
|
1901 | 1898 | |
|
|
1902 | 1899 | __buffer_data = [] |
|
|
1903 | 1900 | __buffer_times = [] |
|
|
1904 | 1901 | |
|
|
1905 | 1902 | buffer = None |
|
|
1906 | 1903 | |
|
|
1907 | 1904 | outliers_IDs_list = [] |
|
|
1908 | 1905 | |
|
|
1909 | 1906 | |
|
|
1910 | 1907 | __slots__ = ('n','navg','profileMargin','thHistOutlier','minHei_idx','maxHei_idx','nHeights', |
|
|
1911 | 1908 | '__dh','first_utcBlock','__profIndex','init_prof','end_prof','lenProfileOut','nChannels', |
|
|
1912 | 1909 | '__count_exec','__initime','__dataReady','__ipp') |
|
|
1913 | 1910 | def __init__(self, **kwargs): |
|
|
1914 | 1911 | |
|
|
1915 | 1912 | Operation.__init__(self, **kwargs) |
|
|
1916 | 1913 | self.isConfig = False |
|
|
1917 | 1914 | |
|
|
1918 | 1915 | def setup(self,dataOut, n=None , navg=0.8, profileMargin=50,thHistOutlier=3, minHei=None, maxHei=None): |
|
|
1919 | 1916 | |
|
|
1920 | 1917 | if n == None and timeInterval == None: |
|
|
1921 | 1918 | raise ValueError("nprofiles or timeInterval should be specified ...") |
|
|
1922 | 1919 | |
|
|
1923 | 1920 | if n != None: |
|
|
1924 | 1921 | self.n = n |
|
|
1925 | 1922 | |
|
|
1926 | 1923 | self.navg = navg |
|
|
1927 | 1924 | self.profileMargin = profileMargin |
|
|
1928 | 1925 | self.thHistOutlier = thHistOutlier |
|
|
1929 | 1926 | self.__profIndex = 0 |
|
|
1930 | 1927 | self.buffer = None |
|
|
1931 | 1928 | self._ipp = dataOut.ippSeconds |
|
|
1932 | 1929 | self.n_prof_released = 0 |
|
|
1933 | 1930 | self.heightList = dataOut.heightList |
|
|
1934 | 1931 | self.init_prof = 0 |
|
|
1935 | 1932 | self.end_prof = 0 |
|
|
1936 | 1933 | self.__count_exec = 0 |
|
|
1937 | 1934 | self.__profIndex = 0 |
|
|
1938 | 1935 | self.first_utcBlock = None |
|
|
1939 | 1936 | self.__dh = dataOut.heightList[1] - dataOut.heightList[0] |
|
|
1940 | 1937 | minHei = minHei |
|
|
1941 | 1938 | maxHei = maxHei |
|
|
1942 | 1939 | if minHei==None : |
|
|
1943 | 1940 | minHei = dataOut.heightList[0] |
|
|
1944 | 1941 | if maxHei==None : |
|
|
1945 | 1942 | maxHei = dataOut.heightList[-1] |
|
|
1946 | 1943 | self.minHei_idx,self.maxHei_idx = getHei_index(minHei, maxHei, dataOut.heightList) |
|
|
1947 | 1944 | |
|
|
1948 | 1945 | self.nChannels = dataOut.nChannels |
|
|
1949 | 1946 | self.nHeights = dataOut.nHeights |
|
|
1950 | 1947 | |
|
|
1951 | 1948 | def filterSatsProfiles(self): |
|
|
1952 | 1949 | data = self.__buffer_data |
|
|
1953 | 1950 | #print(data.shape) |
|
|
1954 | 1951 | nChannels, profiles, heights = data.shape |
|
|
1955 | 1952 | indexes=[] |
|
|
1956 | 1953 | outliers_IDs=[] |
|
|
1957 | 1954 | for c in range(nChannels): |
|
|
1958 | 1955 | for h in range(self.minHei_idx, self.maxHei_idx): |
|
|
1959 | 1956 | power = data[c,:,h] * numpy.conjugate(data[c,:,h]) |
|
|
1960 | 1957 | power = power.real |
|
|
1961 | 1958 | #power = (numpy.abs(data[c,:,h].real)) |
|
|
1962 | 1959 | sortdata = numpy.sort(power, axis=None) |
|
|
1963 | 1960 | sortID=power.argsort() |
|
|
1964 | 1961 | index = _noise.hildebrand_sekhon2(sortdata,self.navg) #0.75-> buen valor |
|
|
1965 | 1962 | |
|
|
1966 | 1963 | indexes.append(index) |
|
|
1967 | 1964 | outliers_IDs=numpy.append(outliers_IDs,sortID[index:]) |
|
|
1968 | 1965 | # print(outliers_IDs) |
|
|
1969 | 1966 | # fig,ax = plt.subplots() |
|
|
1970 | 1967 | # #ax.set_title(str(k)+" "+str(j)) |
|
|
1971 | 1968 | # x=range(len(sortdata)) |
|
|
1972 | 1969 | # ax.scatter(x,sortdata) |
|
|
1973 | 1970 | # ax.axvline(index) |
|
|
1974 | 1971 | # plt.grid() |
|
|
1975 | 1972 | # plt.show() |
|
|
1976 | 1973 | |
|
|
1977 | 1974 | |
|
|
1978 | 1975 | outliers_IDs = outliers_IDs.astype(numpy.dtype('int64')) |
|
|
1979 | 1976 | outliers_IDs = numpy.unique(outliers_IDs) |
|
|
1980 | 1977 | outs_lines = numpy.sort(outliers_IDs) |
|
|
1981 | 1978 | # #print("outliers Ids: ", outs_lines, outs_lines.shape) |
|
|
1982 | 1979 | #hist, bin_edges = numpy.histogram(outs_lines, bins=10, density=True) |
|
|
1983 | 1980 | |
|
|
1984 | 1981 | |
|
|
1985 | 1982 | #Agrupando el histograma de outliers, |
|
|
1986 | 1983 | #my_bins = numpy.linspace(0,int(profiles), int(profiles/100), endpoint=False) |
|
|
1987 | 1984 | my_bins = numpy.linspace(0,9600, 96, endpoint=False) |
|
|
1988 | 1985 | |
|
|
1989 | 1986 | hist, bins = numpy.histogram(outs_lines,bins=my_bins) |
|
|
1990 | 1987 | hist_outliers_indexes = numpy.where(hist > self.thHistOutlier) #es outlier |
|
|
1991 | 1988 | #print(hist_outliers_indexes[0]) |
|
|
1992 | 1989 | bins_outliers_indexes = [int(i) for i in bins[hist_outliers_indexes]] # |
|
|
1993 | 1990 | #print(bins_outliers_indexes) |
|
|
1994 | 1991 | outlier_loc_index = [] |
|
|
1995 | 1992 | |
|
|
1996 | 1993 | |
|
|
1997 | 1994 | # for n in range(len(bins_outliers_indexes)-1): |
|
|
1998 | 1995 | # for k in range(bins_outliers_indexes[n]-self.profileMargin,bins_outliers_indexes[n+1]+self.profileMargin): |
|
|
1999 | 1996 | # outlier_loc_index.append(k) |
|
|
2000 | 1997 | |
|
|
2001 | 1998 | outlier_loc_index = [e for n in range(len(bins_outliers_indexes)-1) for e in range(bins_outliers_indexes[n]-self.profileMargin,bins_outliers_indexes[n+1]+self.profileMargin) ] |
|
|
2002 | 1999 | |
|
|
2003 | 2000 | outlier_loc_index = numpy.asarray(outlier_loc_index) |
|
|
2004 | 2001 | #print(len(numpy.unique(outlier_loc_index)), numpy.unique(outlier_loc_index)) |
|
|
2005 | 2002 | |
|
|
2006 | 2003 | |
|
|
2007 | 2004 | |
|
|
2008 | 2005 | # x, y = numpy.meshgrid(numpy.arange(profiles), self.heightList) |
|
|
2009 | 2006 | # fig, ax = plt.subplots(1,2,figsize=(8, 6)) |
|
|
2010 | 2007 | # |
|
|
2011 | 2008 | # dat = data[0,:,:].real |
|
|
2012 | 2009 | # m = numpy.nanmean(dat) |
|
|
2013 | 2010 | # o = numpy.nanstd(dat) |
|
|
2014 | 2011 | # #print(m, o, x.shape, y.shape) |
|
|
2015 | 2012 | # c = ax[0].pcolormesh(x, y, dat.T, cmap ='YlGnBu', vmin = (m-2*o), vmax = (m+2*o)) |
|
|
2016 | 2013 | # ax[0].vlines(outs_lines,200,600, linestyles='dashed', label = 'outs', color='w') |
|
|
2017 | 2014 | # fig.colorbar(c) |
|
|
2018 | 2015 | # ax[0].vlines(outlier_loc_index,650,750, linestyles='dashed', label = 'outs', color='r') |
|
|
2019 | 2016 | # ax[1].hist(outs_lines,bins=my_bins) |
|
|
2020 | 2017 | # plt.show() |
|
|
2021 | 2018 | |
|
|
2022 | 2019 | |
|
|
2023 | 2020 | self.outliers_IDs_list = numpy.unique(outlier_loc_index) |
|
|
2024 | 2021 | return data |
|
|
2025 | 2022 | |
|
|
2026 | 2023 | def cleanOutliersByBlock(self): |
|
|
2024 | import matplotlib.pyplot as plt | |
|
|
2025 | import datetime | |
|
|
2027 | 2026 | #print(self.__buffer_data[0].shape) |
|
|
2028 | 2027 | data = self.__buffer_data#.copy() |
|
|
2029 |
|
|
|
|
2030 | ''' | |
|
|
2028 | print("cleaning shape inpt: ",data.shape) | |
|
|
2029 | ||
|
|
2031 | 2030 | self.__buffer_data = [] |
|
|
2032 | 2031 | |
|
|
2033 | spectrum = numpy.fft.fft2(data, axes=(0,2)) | |
|
|
2034 |
|
|
|
|
2032 | spectrum = numpy.fft.fft2(data[:,:,self.minHei_idx:], axes=(0,2)) | |
|
|
2033 | print("spc : ",spectrum.shape) | |
|
|
2035 | 2034 | (nch,nsamples, nh) = spectrum.shape |
|
|
2036 | 2035 | data2 = None |
|
|
2037 | 2036 | #print(data.shape) |
|
|
2038 | 2037 | spectrum2 = spectrum.copy() |
|
|
2039 | 2038 | for ch in range(nch): |
|
|
2040 | 2039 | dh = self.__dh |
|
|
2041 | 2040 | dt1 = (dh*1.0e-6)/(0.15) |
|
|
2042 | 2041 | dt2 = self.__buffer_times[1]-self.__buffer_times[0] |
|
|
2043 | 2042 | freqv = numpy.fft.fftfreq(nh, d=dt1) |
|
|
2044 | 2043 | freqh = numpy.fft.fftfreq(self.n, d=dt2) |
|
|
2045 | 2044 | #print("spc loop: ") |
|
|
2046 | 2045 | |
|
|
2047 | 2046 | |
|
|
2048 | 2047 | |
|
|
2049 | 2048 | x, y = numpy.meshgrid(numpy.sort(freqh),numpy.sort(freqv)) |
|
|
2050 | 2049 | z = numpy.abs(spectrum[ch,:,:]) |
|
|
2050 | phase = numpy.angle(spectrum[ch,:,:]) | |
|
|
2051 | 2051 | # Find all peaks higher than the 98th percentile |
|
|
2052 |
peaks = z < numpy.percentile(z, 9 |
|
|
|
2052 | peaks = z < numpy.percentile(z, 99) | |
|
|
2053 | 2053 | #print(peaks) |
|
|
2054 | 2054 | # Set those peak coefficients to zero |
|
|
2055 | 2055 | spectrum2 = spectrum2 * peaks.astype(int) |
|
|
2056 | data2 = numpy.fft.ifft2(spectrum2) | |
|
|
2056 | data2 = numpy.fft.ifft2(spectrum2,axes=(0,2)) | |
|
|
2057 | 2057 | |
|
|
2058 | 2058 | dat = numpy.log10(z.T) |
|
|
2059 |
dat |
|
|
|
2060 | ||
|
|
2061 | # m = numpy.mean(dat) | |
|
|
2062 |
|
|
|
|
2063 | # fig, ax = plt.subplots(2,1,figsize=(8, 6)) | |
|
|
2064 | # | |
|
|
2065 | # c = ax[0].pcolormesh(x, y, dat, cmap ='YlGnBu', vmin = (m-2*o), vmax = (m+2*o)) | |
|
|
2066 | # #c = ax.pcolor( z.T , cmap ='gray', vmin = (m-2*o), vmax = (m+2*o)) | |
|
|
2067 | # date_time = datetime.datetime.fromtimestamp(self.__buffer_times[0]).strftime('%Y-%m-%d %H:%M:%S.%f') | |
|
|
2068 | # #strftime('%Y-%m-%d %H:%M:%S') | |
|
|
2069 | # ax[0].set_title('Spectrum magnitude '+date_time) | |
|
|
2070 | # fig.canvas.set_window_title('Spectrum magnitude {} '.format(self.n)+date_time) | |
|
|
2071 | # | |
|
|
2072 | # | |
|
|
2073 | # c = ax[1].pcolormesh(x, y, dat, cmap ='YlGnBu', vmin = (m-2*o), vmax = (m+2*o)) | |
|
|
2074 | # fig.colorbar(c) | |
|
|
2075 | # plt.show() | |
|
|
2059 | pdat = numpy.log10(phase.T) | |
|
|
2060 | dat2 = numpy.log10(numpy.abs(spectrum2.T)) | |
|
|
2061 | ||
|
|
2062 | m = numpy.mean(dat) | |
|
|
2063 | o = numpy.std(dat) | |
|
|
2064 | fig, ax = plt.subplots(1,2,figsize=(12, 6)) | |
|
|
2065 | ||
|
|
2066 | colormap = 'jet' | |
|
|
2067 | #c = ax[0].pcolormesh(x, y, dat, cmap =colormap, vmin = (m-2*o)/2, vmax = (m+2*o)) | |
|
|
2068 | c = ax[0].pcolormesh(x, y, dat, cmap =colormap, vmin = 4.2, vmax = 5.0) | |
|
|
2069 | fig.colorbar(c, ax=ax[0]) | |
|
|
2070 | #print("aqui estoy", dat.shape) | |
|
|
2071 | #c = ax.pcolor( z.T , cmap ='gray', vmin = (m-2*o), vmax = (m+2*o)) | |
|
|
2072 | date_time = datetime.datetime.fromtimestamp(self.__buffer_times[0]).strftime('%Y-%m-%d %H:%M:%S.%f') | |
|
|
2073 | #strftime('%Y-%m-%d %H:%M:%S') | |
|
|
2074 | #ax[0].set_title('Spectrum magnitude '+date_time) | |
|
|
2075 | fig.canvas.set_window_title('Spectrum magnitude {} '.format(self.n)+date_time) | |
|
|
2076 | #print("aqui estoy2",dat2[:,:,0].shape) | |
|
|
2077 | c = ax[1].pcolormesh(x, y, pdat, cmap =colormap, vmin = -0.0, vmax = 0.5) | |
|
|
2078 | #c = ax[1].pcolormesh(x, y, dat2[:,:,0], cmap =colormap, vmin = (m-2*o)/2, vmax = (m+2*o)-1) | |
|
|
2079 | #print("aqui estoy3") | |
|
|
2080 | fig.colorbar(c, ax=ax[1]) | |
|
|
2081 | plt.show() | |
|
|
2076 | 2082 | |
|
|
2077 | 2083 | #print(data2.shape) |
|
|
2078 | 2084 | |
|
|
2079 | data = data2 | |
|
|
2085 | #data = data2 | |
|
|
2086 | ||
|
|
2087 | #cleanBlock = numpy.fft.ifft2(data, axes=(0,2)).reshape() | |
|
|
2080 | 2088 | |
|
|
2081 | #cleanBlock = numpy.fft.ifft2(spectrum, axes=(0,2)).reshape() | |
|
|
2082 | ''' | |
|
|
2083 | 2089 | #print("cleanOutliersByBlock Done") |
|
|
2084 | 2090 | |
|
|
2085 | return self.filterSatsProfiles() | |
|
|
2091 | return data | |
|
|
2086 | 2092 | |
|
|
2087 | 2093 | |
|
|
2088 | 2094 | |
|
|
2089 | 2095 | def fillBuffer(self, data, datatime): |
|
|
2090 | 2096 | |
|
|
2091 | 2097 | if self.__profIndex == 0: |
|
|
2092 | 2098 | self.__buffer_data = data.copy() |
|
|
2093 | 2099 | |
|
|
2094 | 2100 | else: |
|
|
2095 | 2101 | self.__buffer_data = numpy.concatenate((self.__buffer_data,data), axis=1)#en perfiles |
|
|
2096 | 2102 | self.__profIndex += 1 |
|
|
2097 |
|
|
|
|
2103 | self.__buffer_times.append(datatime) | |
|
|
2098 | 2104 | |
|
|
2099 | 2105 | def getData(self, data, datatime=None): |
|
|
2100 | 2106 | |
|
|
2101 | 2107 | if self.__profIndex == 0: |
|
|
2102 | 2108 | self.__initime = datatime |
|
|
2103 | 2109 | |
|
|
2104 | 2110 | |
|
|
2105 | 2111 | self.__dataReady = False |
|
|
2106 | 2112 | |
|
|
2107 | 2113 | self.fillBuffer(data, datatime) |
|
|
2108 | 2114 | dataBlock = None |
|
|
2109 | 2115 | |
|
|
2110 | 2116 | if self.__profIndex == self.n: |
|
|
2111 | 2117 | #print("apnd : ",data) |
|
|
2112 | 2118 | #dataBlock = self.cleanOutliersByBlock() |
|
|
2113 | 2119 | dataBlock = self.filterSatsProfiles() |
|
|
2114 | 2120 | self.__dataReady = True |
|
|
2115 | 2121 | |
|
|
2116 | 2122 | return dataBlock |
|
|
2117 | 2123 | |
|
|
2118 | 2124 | if dataBlock is None: |
|
|
2119 | 2125 | return None, None |
|
|
2120 | 2126 | |
|
|
2121 | 2127 | |
|
|
2122 | 2128 | |
|
|
2123 | 2129 | return dataBlock |
|
|
2124 | 2130 | |
|
|
2125 | 2131 | def releaseBlock(self): |
|
|
2126 | 2132 | |
|
|
2127 | 2133 | if self.n % self.lenProfileOut != 0: |
|
|
2128 | 2134 | raise ValueError("lenProfileOut %d must be submultiple of nProfiles %d" %(self.lenProfileOut, self.n)) |
|
|
2129 | 2135 | return None |
|
|
2130 | 2136 | |
|
|
2131 | 2137 | data = self.buffer[:,self.init_prof:self.end_prof:,:] #ch, prof, alt |
|
|
2132 | 2138 | |
|
|
2133 | 2139 | self.init_prof = self.end_prof |
|
|
2134 | 2140 | self.end_prof += self.lenProfileOut |
|
|
2135 | 2141 | #print("data release shape: ",dataOut.data.shape, self.end_prof) |
|
|
2136 | 2142 | self.n_prof_released += 1 |
|
|
2137 | 2143 | |
|
|
2138 | 2144 | |
|
|
2139 | 2145 | #print("f_no_data ", dataOut.flagNoData) |
|
|
2140 | 2146 | return data |
|
|
2141 | 2147 | |
|
|
2142 | 2148 | def run(self, dataOut, n=None, navg=0.8, nProfilesOut=1, profile_margin=50,th_hist_outlier=3,minHei=None, maxHei=None): |
|
|
2143 | 2149 | #print("run op buffer 2D",dataOut.ippSeconds) |
|
|
2144 | 2150 | # self.nChannels = dataOut.nChannels |
|
|
2145 | 2151 | # self.nHeights = dataOut.nHeights |
|
|
2146 | 2152 | |
|
|
2147 | 2153 | if not self.isConfig: |
|
|
2148 | 2154 | #print("init p idx: ", dataOut.profileIndex ) |
|
|
2149 | 2155 | self.setup(dataOut,n=n, navg=navg,profileMargin=profile_margin, |
|
|
2150 | 2156 | thHistOutlier=th_hist_outlier,minHei=minHei, maxHei=maxHei) |
|
|
2151 | 2157 | self.isConfig = True |
|
|
2152 | 2158 | |
|
|
2153 | 2159 | dataBlock = None |
|
|
2154 | 2160 | |
|
|
2155 | 2161 | if not dataOut.buffer_empty: #hay datos acumulados |
|
|
2156 | 2162 | |
|
|
2157 | 2163 | if self.init_prof == 0: |
|
|
2158 | 2164 | self.n_prof_released = 0 |
|
|
2159 | 2165 | self.lenProfileOut = nProfilesOut |
|
|
2160 | 2166 | dataOut.flagNoData = False |
|
|
2161 | 2167 | #print("tp 2 ",dataOut.data.shape) |
|
|
2162 | 2168 | |
|
|
2163 | 2169 | self.init_prof = 0 |
|
|
2164 | 2170 | self.end_prof = self.lenProfileOut |
|
|
2165 | 2171 | |
|
|
2166 | 2172 | dataOut.nProfiles = self.lenProfileOut |
|
|
2167 | 2173 | if nProfilesOut == 1: |
|
|
2168 | 2174 | dataOut.flagDataAsBlock = False |
|
|
2169 | 2175 | else: |
|
|
2170 | 2176 | dataOut.flagDataAsBlock = True |
|
|
2171 | 2177 | #print("prof: ",self.init_prof) |
|
|
2172 | 2178 | dataOut.flagNoData = False |
|
|
2173 | 2179 | if numpy.isin(self.n_prof_released, self.outliers_IDs_list): |
|
|
2174 | 2180 | #print("omitting: ", self.n_prof_released) |
|
|
2175 | 2181 | dataOut.flagNoData = True |
|
|
2176 | 2182 | dataOut.ippSeconds = self._ipp |
|
|
2177 | 2183 | dataOut.utctime = self.first_utcBlock + self.init_prof*self._ipp |
|
|
2178 | 2184 | # print("time: ", dataOut.utctime, self.first_utcBlock, self.init_prof,self._ipp,dataOut.ippSeconds) |
|
|
2179 | 2185 | #dataOut.data = self.releaseBlock() |
|
|
2180 | 2186 | #########################################################3 |
|
|
2181 | 2187 | if self.n % self.lenProfileOut != 0: |
|
|
2182 | 2188 | raise ValueError("lenProfileOut %d must be submultiple of nProfiles %d" %(self.lenProfileOut, self.n)) |
|
|
2183 | 2189 | return None |
|
|
2184 | 2190 | |
|
|
2185 | 2191 | dataOut.data = self.buffer[:,self.init_prof:self.end_prof:,:] #ch, prof, alt |
|
|
2186 | 2192 | |
|
|
2187 | 2193 | self.init_prof = self.end_prof |
|
|
2188 | 2194 | self.end_prof += self.lenProfileOut |
|
|
2189 | 2195 | #print("data release shape: ",dataOut.data.shape, self.end_prof) |
|
|
2190 | 2196 | self.n_prof_released += 1 |
|
|
2191 | 2197 | |
|
|
2192 | 2198 | if self.end_prof >= (self.n +self.lenProfileOut): |
|
|
2193 | 2199 | |
|
|
2194 | 2200 | self.init_prof = 0 |
|
|
2195 | 2201 | self.__profIndex = 0 |
|
|
2196 | 2202 | self.buffer = None |
|
|
2197 | 2203 | dataOut.buffer_empty = True |
|
|
2198 | 2204 | self.outliers_IDs_list = [] |
|
|
2199 | 2205 | self.n_prof_released = 0 |
|
|
2200 | 2206 | dataOut.flagNoData = False #enviar ultimo aunque sea outlier :( |
|
|
2201 | 2207 | #print("cleaning...", dataOut.buffer_empty) |
|
|
2202 | 2208 | dataOut.profileIndex = 0 #self.lenProfileOut |
|
|
2203 | 2209 | #################################################################### |
|
|
2204 | 2210 | return dataOut |
|
|
2205 | 2211 | |
|
|
2206 | 2212 | |
|
|
2207 | 2213 | #print("tp 223 ",dataOut.data.shape) |
|
|
2208 | 2214 | dataOut.flagNoData = True |
|
|
2209 | 2215 | |
|
|
2210 | 2216 | |
|
|
2211 | 2217 | |
|
|
2212 | 2218 | try: |
|
|
2213 | 2219 | #dataBlock = self.getData(dataOut.data.reshape(self.nChannels,1,self.nHeights), dataOut.utctime) |
|
|
2214 | 2220 | dataBlock = self.getData(numpy.reshape(dataOut.data,(self.nChannels,1,self.nHeights)), dataOut.utctime) |
|
|
2215 | 2221 | self.__count_exec +=1 |
|
|
2216 | 2222 | except Exception as e: |
|
|
2217 | 2223 | print("Error getting profiles data",self.__count_exec ) |
|
|
2218 | 2224 | print(e) |
|
|
2219 | 2225 | sys.exit() |
|
|
2220 | 2226 | |
|
|
2221 | 2227 | if self.__dataReady: |
|
|
2222 | 2228 | #print("omitting: ", len(self.outliers_IDs_list)) |
|
|
2223 | 2229 | self.__count_exec = 0 |
|
|
2224 | 2230 | #dataOut.data = |
|
|
2225 | 2231 | #self.buffer = numpy.flip(dataBlock, axis=1) |
|
|
2226 | 2232 | self.buffer = dataBlock |
|
|
2227 | 2233 | self.first_utcBlock = self.__initime |
|
|
2228 | 2234 | dataOut.utctime = self.__initime |
|
|
2229 | 2235 | dataOut.nProfiles = self.__profIndex |
|
|
2230 | 2236 | #dataOut.flagNoData = False |
|
|
2231 | 2237 | self.init_prof = 0 |
|
|
2232 | 2238 | self.__profIndex = 0 |
|
|
2233 | 2239 | self.__initime = None |
|
|
2234 | 2240 | dataBlock = None |
|
|
2235 | 2241 | self.__buffer_times = [] |
|
|
2236 | 2242 | dataOut.error = False |
|
|
2237 | 2243 | dataOut.useInputBuffer = True |
|
|
2238 | 2244 | dataOut.buffer_empty = False |
|
|
2239 | 2245 | #print("1 ch: {} prof: {} hs: {}".format(int(dataOut.nChannels),int(dataOut.nProfiles),int(dataOut.nHeights))) |
|
|
2240 | 2246 | |
|
|
2241 | 2247 | |
|
|
2242 | 2248 | |
|
|
2243 | 2249 | #print(self.__count_exec) |
|
|
2244 | 2250 | |
|
|
2245 | 2251 | return dataOut |
|
|
2246 | 2252 | |
|
|
2247 | 2253 | class RemoveProfileSats(Operation): |
|
|
2248 | 2254 | ''' |
|
|
2249 | 2255 | Omite los perfiles contaminados con seΓ±al de satelites, |
|
|
2250 | 2256 | In: minHei = min_sat_range |
|
|
2251 | 2257 | max_sat_range |
|
|
2252 | 2258 | min_hei_ref |
|
|
2253 | 2259 | max_hei_ref |
|
|
2254 | 2260 | th = diference between profiles mean, ref and sats |
|
|
2255 | 2261 | Out: |
|
|
2256 | 2262 | profile clean |
|
|
2257 | 2263 | ''' |
|
|
2258 | 2264 | |
|
|
2259 | 2265 | isConfig = False |
|
|
2260 | 2266 | min_sats = 0 |
|
|
2261 | 2267 | max_sats = 999999999 |
|
|
2262 | 2268 | min_ref= 0 |
|
|
2263 | 2269 | max_ref= 9999999999 |
|
|
2264 | 2270 | needReshape = False |
|
|
2265 | 2271 | count = 0 |
|
|
2266 | 2272 | thdB = 0 |
|
|
2267 | 2273 | byRanges = False |
|
|
2268 | 2274 | min_sats = None |
|
|
2269 | 2275 | max_sats = None |
|
|
2270 | 2276 | noise = 0 |
|
|
2271 | 2277 | |
|
|
2272 | 2278 | def __init__(self, **kwargs): |
|
|
2273 | 2279 | |
|
|
2274 | 2280 | Operation.__init__(self, **kwargs) |
|
|
2275 | 2281 | self.isConfig = False |
|
|
2276 | 2282 | |
|
|
2277 | 2283 | |
|
|
2278 | 2284 | def setup(self, dataOut, minHei, maxHei, minRef, maxRef, th, thdB, rangeHeiList): |
|
|
2279 | 2285 | |
|
|
2280 | 2286 | if rangeHeiList!=None: |
|
|
2281 | 2287 | self.byRanges = True |
|
|
2282 | 2288 | else: |
|
|
2283 | 2289 | if minHei==None or maxHei==None : |
|
|
2284 | 2290 | raise ValueError("Parameters heights are required") |
|
|
2285 | 2291 | if minRef==None or maxRef==None: |
|
|
2286 | 2292 | raise ValueError("Parameters heights are required") |
|
|
2287 | 2293 | |
|
|
2288 | 2294 | if self.byRanges: |
|
|
2289 | 2295 | self.min_sats = [] |
|
|
2290 | 2296 | self.max_sats = [] |
|
|
2291 | 2297 | for min,max in rangeHeiList: |
|
|
2292 | 2298 | a,b = getHei_index(min, max, dataOut.heightList) |
|
|
2293 | 2299 | self.min_sats.append(a) |
|
|
2294 | 2300 | self.max_sats.append(b) |
|
|
2295 | 2301 | else: |
|
|
2296 | 2302 | self.min_sats, self.max_sats = getHei_index(minHei, maxHei, dataOut.heightList) |
|
|
2297 | 2303 | self.min_ref, self.max_ref = getHei_index(minRef, maxRef, dataOut.heightList) |
|
|
2298 | 2304 | self.th = th |
|
|
2299 | 2305 | self.thdB = thdB |
|
|
2300 | 2306 | self.isConfig = True |
|
|
2301 | 2307 | |
|
|
2302 | 2308 | |
|
|
2303 | 2309 | def compareRanges(self,data, minHei,maxHei): |
|
|
2304 | 2310 | |
|
|
2305 | 2311 | # ref = data[0,self.min_ref:self.max_ref] * numpy.conjugate(data[0,self.min_ref:self.max_ref]) |
|
|
2306 | 2312 | # p_ref = 10*numpy.log10(ref.real) |
|
|
2307 | 2313 | # m_ref = numpy.mean(p_ref) |
|
|
2308 | 2314 | |
|
|
2309 | 2315 | m_ref = self.noise |
|
|
2310 | 2316 | |
|
|
2311 | 2317 | sats = data[0,minHei:maxHei] * numpy.conjugate(data[0,minHei:maxHei]) |
|
|
2312 | 2318 | p_sats = 10*numpy.log10(sats.real) |
|
|
2313 | 2319 | m_sats = numpy.mean(p_sats) |
|
|
2314 | 2320 | |
|
|
2315 | 2321 | if m_sats > (m_ref + self.th): #and (m_sats > self.thdB): |
|
|
2316 | 2322 | #print("msats: ",m_sats," \tmRef: ", m_ref, "\t",(m_sats - m_ref)) |
|
|
2317 | 2323 | #print("Removing profiles...") |
|
|
2318 | 2324 | return False |
|
|
2319 | 2325 | |
|
|
2320 | 2326 | return True |
|
|
2321 | 2327 | |
|
|
2322 | 2328 | def isProfileClean(self, data): |
|
|
2323 | 2329 | ''' |
|
|
2324 | 2330 | Analiza solo 1 canal, y descarta todos... |
|
|
2325 | 2331 | ''' |
|
|
2326 | 2332 | |
|
|
2327 | 2333 | clean = True |
|
|
2328 | 2334 | |
|
|
2329 | 2335 | if self.byRanges: |
|
|
2330 | 2336 | |
|
|
2331 | 2337 | for n in range(len(self.min_sats)): |
|
|
2332 | 2338 | c = self.compareRanges(data,self.min_sats[n],self.max_sats[n]) |
|
|
2333 | 2339 | clean = clean and c |
|
|
2334 | 2340 | else: |
|
|
2335 | 2341 | |
|
|
2336 | 2342 | clean = (self.compareRanges(data, self.min_sats,self.max_sats)) |
|
|
2337 | 2343 | |
|
|
2338 | 2344 | return clean |
|
|
2339 | 2345 | |
|
|
2340 | 2346 | |
|
|
2341 | 2347 | |
|
|
2342 | 2348 | def run(self, dataOut, minHei=None, maxHei=None, minRef=None, maxRef=None, th=5, thdB=65, rangeHeiList=None): |
|
|
2343 | 2349 | dataOut.flagNoData = True |
|
|
2344 | 2350 | |
|
|
2345 | 2351 | if not self.isConfig: |
|
|
2346 | 2352 | self.setup(dataOut, minHei, maxHei, minRef, maxRef, th, thdB, rangeHeiList) |
|
|
2347 | 2353 | self.isConfig = True |
|
|
2348 | 2354 | #print(self.min_sats,self.max_sats) |
|
|
2349 | 2355 | if dataOut.flagDataAsBlock: |
|
|
2350 | 2356 | raise ValueError("ProfileConcat can only be used when voltage have been read profile by profile, getBlock = False") |
|
|
2351 | 2357 | |
|
|
2352 | 2358 | else: |
|
|
2353 | 2359 | self.noise =10*numpy.log10(dataOut.getNoisebyHildebrand(ymin_index=self.min_ref, ymax_index=self.max_ref)) |
|
|
2354 | 2360 | if not self.isProfileClean(dataOut.data): |
|
|
2355 | 2361 | return dataOut |
|
|
2356 | 2362 | #dataOut.data = numpy.full((dataOut.nChannels,dataOut.nHeights),numpy.NAN) |
|
|
2357 | 2363 | #self.count += 1 |
|
|
2358 | 2364 | |
|
|
2359 | 2365 | dataOut.flagNoData = False |
|
|
2360 | 2366 | |
|
|
2361 | 2367 | return dataOut |
General Comments 0
You need to be logged in to leave comments.
Login now