In [33]:
%matplotlib inline
from matplotlib import pylab as pl
import cPickle as pickle
import pandas as pd
import numpy as np
import os
import random
In [34]:
import sys
sys.path.append('..')
In [35]:
target = 'Dog_1'
data_type = 'interictal' # preictal interictal, test
according to 140905-feature-importance the most important cannel is
In [36]:
channel = 3
get the first segment chain
In [46]:
import scipy.io
nchain = 0 # which chain you want
all_data = last_sequence = last_data_length_sec = last_Fs = last_channels = last_d_shape = None
for segment in range(10000):
fname = '../seizure-data/%s/%s_%s_segment_%04d.mat'%(target,target,data_type,segment+1)
try:
data = scipy.io.loadmat(fname)
except:
break
k = '%s_segment_%d'%(data_type,segment+1)
data_length_sec = data[k]['data_length_sec'][0,0][0,0]
sequence = data[k]['sequence'][0,0][0,0]
Fs = float(data[k]['sampling_frequency'][0,0][0,0])
channels = [t[0] for t in data[k]['channels'][0,0][0]]
d = data[k]['data'][0,0]
assert len(channels) == d.shape[0]
assert int(Fs*data_length_sec + 0.5) == d.shape[1],int(Fs*data_length_sec + 0.5)
assert last_data_length_sec is None or last_data_length_sec == data_length_sec
last_data_length_sec = data_length_sec
assert last_Fs is None or last_Fs == Fs
last_Fs = Fs
assert last_channels is None or all(c1==c2 for c1,c2 in zip(last_channels, channels))
last_channels = channels
assert last_d_shape is None or last_d_shape == d.shape
last_d_shape = d.shape
if last_sequence is None:
all_data = d.astype(float)
elif last_sequence < sequence:
all_data = np.hstack((all_data,d.astype(float)))
else:
if nchain == 0:
break
nchain -= 1
all_data = d.astype(float)
last_sequence = sequence
all_data = all_data.copy()
data_length_sec, sequence, Fs, all_data.shape
Out[46]:
In [47]:
print 'next segment', segment, last_sequence
In [48]:
Nchannels = 4 # d.shape[0]
for i in range(Nchannels):
pl.subplot(Nchannels,1,i+1)
pl.plot(all_data[i,:500000])
pl.title(channels[i])
pl.gcf().set_size_inches(14,8);
In [49]:
Y = np.abs(np.fft.rfft(all_data[channel,:]))
In [50]:
Y[60*data_length_sec]/np.std(all_data[0,:])
Out[50]:
In [51]:
Fmax = Fs/2
pl.plot(np.linspace(0,Fmax,Fmax*data_length_sec),Y[:Fmax*data_length_sec])
Out[51]:
In [60]:
NFFT=1024
S = 1438596//6
Pxx, freqs, bins, im = pl.specgram(all_data[0,:], NFFT=NFFT, Fs=Fs, pad_to=NFFT*2, noverlap=NFFT/2)
#pl.gca().set_ylim((0,200.))
pl.gcf().set_size_inches(18,8);
#pl.tight_layout();
pl.autoscale(tight=True);
all_data.shape
Out[60]:
In [58]:
pl.plot(freqs,Pxx.mean(axis=-1))
Out[58]:
In [44]: