|
|
'''
|
|
|
Created on Oct 24, 2016
|
|
|
|
|
|
@author: roj- LouVD
|
|
|
'''
|
|
|
|
|
|
import numpy
|
|
|
import copy
|
|
|
import datetime
|
|
|
import time
|
|
|
from time import gmtime
|
|
|
|
|
|
from numpy import transpose
|
|
|
|
|
|
from .jroproc_base import ProcessingUnit, Operation
|
|
|
from schainpy.model.data.jrodata import Parameters
|
|
|
|
|
|
|
|
|
class BLTRParametersProc(ProcessingUnit):
|
|
|
'''
|
|
|
Processing unit for BLTR parameters data (winds)
|
|
|
|
|
|
Inputs:
|
|
|
self.dataOut.nmodes - Number of operation modes
|
|
|
self.dataOut.nchannels - Number of channels
|
|
|
self.dataOut.nranges - Number of ranges
|
|
|
|
|
|
self.dataOut.data_SNR - SNR array
|
|
|
self.dataOut.data_output - Zonal, Vertical and Meridional velocity array
|
|
|
self.dataOut.height - Height array (km)
|
|
|
self.dataOut.time - Time array (seconds)
|
|
|
|
|
|
self.dataOut.fileIndex -Index of the file currently read
|
|
|
self.dataOut.lat - Latitude coordinate of BLTR location
|
|
|
|
|
|
self.dataOut.doy - Experiment doy (number of the day in the current year)
|
|
|
self.dataOut.month - Experiment month
|
|
|
self.dataOut.day - Experiment day
|
|
|
self.dataOut.year - Experiment year
|
|
|
'''
|
|
|
|
|
|
def __init__(self, **kwargs):
|
|
|
'''
|
|
|
Inputs: None
|
|
|
'''
|
|
|
ProcessingUnit.__init__(self, **kwargs)
|
|
|
self.dataOut = Parameters()
|
|
|
self.isConfig = False
|
|
|
|
|
|
def setup(self, mode):
|
|
|
'''
|
|
|
'''
|
|
|
self.dataOut.mode = mode
|
|
|
|
|
|
def run(self, mode, snr_threshold=None):
|
|
|
'''
|
|
|
Inputs:
|
|
|
mode = High resolution (0) or Low resolution (1) data
|
|
|
snr_threshold = snr filter value
|
|
|
'''
|
|
|
|
|
|
if not self.isConfig:
|
|
|
self.setup(mode)
|
|
|
self.isConfig = True
|
|
|
|
|
|
if self.dataIn.type == 'Parameters':
|
|
|
self.dataOut.copy(self.dataIn)
|
|
|
|
|
|
self.dataOut.data_param = self.dataOut.data[mode]
|
|
|
self.dataOut.heightList = self.dataOut.height[0]
|
|
|
self.dataOut.data_SNR = self.dataOut.data_SNR[mode]
|
|
|
|
|
|
if snr_threshold is not None:
|
|
|
SNRavg = numpy.average(self.dataOut.data_SNR, axis=0)
|
|
|
SNRavgdB = 10*numpy.log10(SNRavg)
|
|
|
for i in range(3):
|
|
|
self.dataOut.data_param[i][SNRavgdB <= snr_threshold] = numpy.nan
|
|
|
|
|
|
# TODO
|
|
|
class OutliersFilter(Operation):
|
|
|
|
|
|
def __init__(self, **kwargs):
|
|
|
'''
|
|
|
'''
|
|
|
Operation.__init__(self, **kwargs)
|
|
|
|
|
|
def run(self, svalue2, method, factor, filter, npoints=9):
|
|
|
'''
|
|
|
Inputs:
|
|
|
svalue - string to select array velocity
|
|
|
svalue2 - string to choose axis filtering
|
|
|
method - 0 for SMOOTH or 1 for MEDIAN
|
|
|
factor - number used to set threshold
|
|
|
filter - 1 for data filtering using the standard deviation criteria else 0
|
|
|
npoints - number of points for mask filter
|
|
|
'''
|
|
|
|
|
|
print(' Outliers Filter {} {} / threshold = {}'.format(svalue, svalue, factor))
|
|
|
|
|
|
|
|
|
yaxis = self.dataOut.heightList
|
|
|
xaxis = numpy.array([[self.dataOut.utctime]])
|
|
|
|
|
|
# Zonal
|
|
|
value_temp = self.dataOut.data_output[0]
|
|
|
|
|
|
# Zonal
|
|
|
value_temp = self.dataOut.data_output[1]
|
|
|
|
|
|
# Vertical
|
|
|
value_temp = numpy.transpose(self.dataOut.data_output[2])
|
|
|
|
|
|
htemp = yaxis
|
|
|
std = value_temp
|
|
|
for h in range(len(htemp)):
|
|
|
nvalues_valid = len(numpy.where(numpy.isfinite(value_temp[h]))[0])
|
|
|
minvalid = npoints
|
|
|
|
|
|
#only if valid values greater than the minimum required (10%)
|
|
|
if nvalues_valid > minvalid:
|
|
|
|
|
|
if method == 0:
|
|
|
#SMOOTH
|
|
|
w = value_temp[h] - self.Smooth(input=value_temp[h], width=npoints, edge_truncate=1)
|
|
|
|
|
|
|
|
|
if method == 1:
|
|
|
#MEDIAN
|
|
|
w = value_temp[h] - self.Median(input=value_temp[h], width = npoints)
|
|
|
|
|
|
dw = numpy.std(w[numpy.where(numpy.isfinite(w))],ddof = 1)
|
|
|
|
|
|
threshold = dw*factor
|
|
|
value_temp[numpy.where(w > threshold),h] = numpy.nan
|
|
|
value_temp[numpy.where(w < -1*threshold),h] = numpy.nan
|
|
|
|
|
|
|
|
|
#At the end
|
|
|
if svalue2 == 'inHeight':
|
|
|
value_temp = numpy.transpose(value_temp)
|
|
|
output_array[:,m] = value_temp
|
|
|
|
|
|
if svalue == 'zonal':
|
|
|
self.dataOut.data_output[0] = output_array
|
|
|
|
|
|
elif svalue == 'meridional':
|
|
|
self.dataOut.data_output[1] = output_array
|
|
|
|
|
|
elif svalue == 'vertical':
|
|
|
self.dataOut.data_output[2] = output_array
|
|
|
|
|
|
return self.dataOut.data_output
|
|
|
|
|
|
|
|
|
def Median(self,input,width):
|
|
|
'''
|
|
|
Inputs:
|
|
|
input - Velocity array
|
|
|
width - Number of points for mask filter
|
|
|
|
|
|
'''
|
|
|
|
|
|
if numpy.mod(width,2) == 1:
|
|
|
pc = int((width - 1) / 2)
|
|
|
cont = 0
|
|
|
output = []
|
|
|
|
|
|
for i in range(len(input)):
|
|
|
if i >= pc and i < len(input) - pc:
|
|
|
new2 = input[i-pc:i+pc+1]
|
|
|
temp = numpy.where(numpy.isfinite(new2))
|
|
|
new = new2[temp]
|
|
|
value = numpy.median(new)
|
|
|
output.append(value)
|
|
|
|
|
|
output = numpy.array(output)
|
|
|
output = numpy.hstack((input[0:pc],output))
|
|
|
output = numpy.hstack((output,input[-pc:len(input)]))
|
|
|
|
|
|
return output
|
|
|
|
|
|
def Smooth(self,input,width,edge_truncate = None):
|
|
|
'''
|
|
|
Inputs:
|
|
|
input - Velocity array
|
|
|
width - Number of points for mask filter
|
|
|
edge_truncate - 1 for truncate the convolution product else
|
|
|
|
|
|
'''
|
|
|
|
|
|
if numpy.mod(width,2) == 0:
|
|
|
real_width = width + 1
|
|
|
nzeros = width / 2
|
|
|
else:
|
|
|
real_width = width
|
|
|
nzeros = (width - 1) / 2
|
|
|
|
|
|
half_width = int(real_width)/2
|
|
|
length = len(input)
|
|
|
|
|
|
gate = numpy.ones(real_width,dtype='float')
|
|
|
norm_of_gate = numpy.sum(gate)
|
|
|
|
|
|
nan_process = 0
|
|
|
nan_id = numpy.where(numpy.isnan(input))
|
|
|
if len(nan_id[0]) > 0:
|
|
|
nan_process = 1
|
|
|
pb = numpy.zeros(len(input))
|
|
|
pb[nan_id] = 1.
|
|
|
input[nan_id] = 0.
|
|
|
|
|
|
if edge_truncate == True:
|
|
|
output = numpy.convolve(input/norm_of_gate,gate,mode='same')
|
|
|
elif edge_truncate == False or edge_truncate == None:
|
|
|
output = numpy.convolve(input/norm_of_gate,gate,mode='valid')
|
|
|
output = numpy.hstack((input[0:half_width],output))
|
|
|
output = numpy.hstack((output,input[len(input)-half_width:len(input)]))
|
|
|
|
|
|
if nan_process:
|
|
|
pb = numpy.convolve(pb/norm_of_gate,gate,mode='valid')
|
|
|
pb = numpy.hstack((numpy.zeros(half_width),pb))
|
|
|
pb = numpy.hstack((pb,numpy.zeros(half_width)))
|
|
|
output[numpy.where(pb > 0.9999)] = numpy.nan
|
|
|
input[nan_id] = numpy.nan
|
|
|
return output
|
|
|
|
|
|
def Average(self,aver=0,nhaver=1):
|
|
|
'''
|
|
|
Inputs:
|
|
|
aver - Indicates the time period over which is averaged or consensus data
|
|
|
nhaver - Indicates the decimation factor in heights
|
|
|
|
|
|
'''
|
|
|
nhpoints = 48
|
|
|
|
|
|
lat_piura = -5.17
|
|
|
lat_huancayo = -12.04
|
|
|
lat_porcuya = -5.8
|
|
|
|
|
|
if '%2.2f'%self.dataOut.lat == '%2.2f'%lat_piura:
|
|
|
hcm = 3.
|
|
|
if self.dataOut.year == 2003 :
|
|
|
if self.dataOut.doy >= 25 and self.dataOut.doy < 64:
|
|
|
nhpoints = 12
|
|
|
|
|
|
elif '%2.2f'%self.dataOut.lat == '%2.2f'%lat_huancayo:
|
|
|
hcm = 3.
|
|
|
if self.dataOut.year == 2003 :
|
|
|
if self.dataOut.doy >= 25 and self.dataOut.doy < 64:
|
|
|
nhpoints = 12
|
|
|
|
|
|
|
|
|
elif '%2.2f'%self.dataOut.lat == '%2.2f'%lat_porcuya:
|
|
|
hcm = 5.#2
|
|
|
|
|
|
pdata = 0.2
|
|
|
taver = [1,2,3,4,6,8,12,24]
|
|
|
t0 = 0
|
|
|
tf = 24
|
|
|
ntime =(tf-t0)/taver[aver]
|
|
|
ti = numpy.arange(ntime)
|
|
|
tf = numpy.arange(ntime) + taver[aver]
|
|
|
|
|
|
|
|
|
old_height = self.dataOut.heightList
|
|
|
|
|
|
if nhaver > 1:
|
|
|
num_hei = len(self.dataOut.heightList)/nhaver/self.dataOut.nmodes
|
|
|
deltha = 0.05*nhaver
|
|
|
minhvalid = pdata*nhaver
|
|
|
for im in range(self.dataOut.nmodes):
|
|
|
new_height = numpy.arange(num_hei)*deltha + self.dataOut.height[im,0] + deltha/2.
|
|
|
|
|
|
|
|
|
data_fHeigths_List = []
|
|
|
data_fZonal_List = []
|
|
|
data_fMeridional_List = []
|
|
|
data_fVertical_List = []
|
|
|
startDTList = []
|
|
|
|
|
|
|
|
|
for i in range(ntime):
|
|
|
height = old_height
|
|
|
|
|
|
start = datetime.datetime(self.dataOut.year,self.dataOut.month,self.dataOut.day) + datetime.timedelta(hours = int(ti[i])) - datetime.timedelta(hours = 5)
|
|
|
stop = datetime.datetime(self.dataOut.year,self.dataOut.month,self.dataOut.day) + datetime.timedelta(hours = int(tf[i])) - datetime.timedelta(hours = 5)
|
|
|
|
|
|
|
|
|
limit_sec1 = time.mktime(start.timetuple())
|
|
|
limit_sec2 = time.mktime(stop.timetuple())
|
|
|
|
|
|
t1 = numpy.where(self.f_timesec >= limit_sec1)
|
|
|
t2 = numpy.where(self.f_timesec < limit_sec2)
|
|
|
time_select = []
|
|
|
for val_sec in t1[0]:
|
|
|
if val_sec in t2[0]:
|
|
|
time_select.append(val_sec)
|
|
|
|
|
|
|
|
|
time_select = numpy.array(time_select,dtype = 'int')
|
|
|
minvalid = numpy.ceil(pdata*nhpoints)
|
|
|
|
|
|
zon_aver = numpy.zeros([self.dataOut.nranges,self.dataOut.nmodes],dtype='f4') + numpy.nan
|
|
|
mer_aver = numpy.zeros([self.dataOut.nranges,self.dataOut.nmodes],dtype='f4') + numpy.nan
|
|
|
ver_aver = numpy.zeros([self.dataOut.nranges,self.dataOut.nmodes],dtype='f4') + numpy.nan
|
|
|
|
|
|
if nhaver > 1:
|
|
|
new_zon_aver = numpy.zeros([num_hei,self.dataOut.nmodes],dtype='f4') + numpy.nan
|
|
|
new_mer_aver = numpy.zeros([num_hei,self.dataOut.nmodes],dtype='f4') + numpy.nan
|
|
|
new_ver_aver = numpy.zeros([num_hei,self.dataOut.nmodes],dtype='f4') + numpy.nan
|
|
|
|
|
|
if len(time_select) > minvalid:
|
|
|
time_average = self.f_timesec[time_select]
|
|
|
|
|
|
for im in range(self.dataOut.nmodes):
|
|
|
|
|
|
for ih in range(self.dataOut.nranges):
|
|
|
if numpy.sum(numpy.isfinite(self.f_zon[time_select,ih,im])) >= minvalid:
|
|
|
zon_aver[ih,im] = numpy.nansum(self.f_zon[time_select,ih,im]) / numpy.sum(numpy.isfinite(self.f_zon[time_select,ih,im]))
|
|
|
|
|
|
if numpy.sum(numpy.isfinite(self.f_mer[time_select,ih,im])) >= minvalid:
|
|
|
mer_aver[ih,im] = numpy.nansum(self.f_mer[time_select,ih,im]) / numpy.sum(numpy.isfinite(self.f_mer[time_select,ih,im]))
|
|
|
|
|
|
if numpy.sum(numpy.isfinite(self.f_ver[time_select,ih,im])) >= minvalid:
|
|
|
ver_aver[ih,im] = numpy.nansum(self.f_ver[time_select,ih,im]) / numpy.sum(numpy.isfinite(self.f_ver[time_select,ih,im]))
|
|
|
|
|
|
if nhaver > 1:
|
|
|
for ih in range(num_hei):
|
|
|
hvalid = numpy.arange(nhaver) + nhaver*ih
|
|
|
|
|
|
if numpy.sum(numpy.isfinite(zon_aver[hvalid,im])) >= minvalid:
|
|
|
new_zon_aver[ih,im] = numpy.nansum(zon_aver[hvalid,im]) / numpy.sum(numpy.isfinite(zon_aver[hvalid,im]))
|
|
|
|
|
|
if numpy.sum(numpy.isfinite(mer_aver[hvalid,im])) >= minvalid:
|
|
|
new_mer_aver[ih,im] = numpy.nansum(mer_aver[hvalid,im]) / numpy.sum(numpy.isfinite(mer_aver[hvalid,im]))
|
|
|
|
|
|
if numpy.sum(numpy.isfinite(ver_aver[hvalid,im])) >= minvalid:
|
|
|
new_ver_aver[ih,im] = numpy.nansum(ver_aver[hvalid,im]) / numpy.sum(numpy.isfinite(ver_aver[hvalid,im]))
|
|
|
if nhaver > 1:
|
|
|
zon_aver = new_zon_aver
|
|
|
mer_aver = new_mer_aver
|
|
|
ver_aver = new_ver_aver
|
|
|
height = new_height
|
|
|
|
|
|
|
|
|
tstart = time_average[0]
|
|
|
tend = time_average[-1]
|
|
|
startTime = time.gmtime(tstart)
|
|
|
|
|
|
year = startTime.tm_year
|
|
|
month = startTime.tm_mon
|
|
|
day = startTime.tm_mday
|
|
|
hour = startTime.tm_hour
|
|
|
minute = startTime.tm_min
|
|
|
second = startTime.tm_sec
|
|
|
|
|
|
startDTList.append(datetime.datetime(year,month,day,hour,minute,second))
|
|
|
|
|
|
|
|
|
o_height = numpy.array([])
|
|
|
o_zon_aver = numpy.array([])
|
|
|
o_mer_aver = numpy.array([])
|
|
|
o_ver_aver = numpy.array([])
|
|
|
if self.dataOut.nmodes > 1:
|
|
|
for im in range(self.dataOut.nmodes):
|
|
|
|
|
|
if im == 0:
|
|
|
h_select = numpy.where(numpy.bitwise_and(height[0,:] >=0,height[0,:] <= hcm,numpy.isfinite(height[0,:])))
|
|
|
else:
|
|
|
h_select = numpy.where(numpy.bitwise_and(height[1,:] > hcm,height[1,:] < 20,numpy.isfinite(height[1,:])))
|
|
|
|
|
|
|
|
|
ht = h_select[0]
|
|
|
|
|
|
o_height = numpy.hstack((o_height,height[im,ht]))
|
|
|
o_zon_aver = numpy.hstack((o_zon_aver,zon_aver[ht,im]))
|
|
|
o_mer_aver = numpy.hstack((o_mer_aver,mer_aver[ht,im]))
|
|
|
o_ver_aver = numpy.hstack((o_ver_aver,ver_aver[ht,im]))
|
|
|
|
|
|
data_fHeigths_List.append(o_height)
|
|
|
data_fZonal_List.append(o_zon_aver)
|
|
|
data_fMeridional_List.append(o_mer_aver)
|
|
|
data_fVertical_List.append(o_ver_aver)
|
|
|
|
|
|
|
|
|
else:
|
|
|
h_select = numpy.where(numpy.bitwise_and(height[0,:] <= hcm,numpy.isfinite(height[0,:])))
|
|
|
ht = h_select[0]
|
|
|
o_height = numpy.hstack((o_height,height[im,ht]))
|
|
|
o_zon_aver = numpy.hstack((o_zon_aver,zon_aver[ht,im]))
|
|
|
o_mer_aver = numpy.hstack((o_mer_aver,mer_aver[ht,im]))
|
|
|
o_ver_aver = numpy.hstack((o_ver_aver,ver_aver[ht,im]))
|
|
|
|
|
|
data_fHeigths_List.append(o_height)
|
|
|
data_fZonal_List.append(o_zon_aver)
|
|
|
data_fMeridional_List.append(o_mer_aver)
|
|
|
data_fVertical_List.append(o_ver_aver)
|
|
|
|
|
|
|
|
|
return startDTList, data_fHeigths_List, data_fZonal_List, data_fMeridional_List, data_fVertical_List
|
|
|
|
|
|
|
|
|
|
