def __init__(self, filename):
self.file = HDFArray(filename)
self.eeg = self.file.getChannelTable('EEG')[1]
self.ecg = self.file.getChannelTable('ECG')[1]
self.fd = self.file.getChannelTable('FD')[1]
self.pnvm = self.file.getChannelTable('PNVM')[1]
self.track = self.file.getChannelTable('TRACK_MARKERS')[1]
self.count_stimuls = 0
self.count_click = 0
for i in range(len(self.track)):
if (self.track[i][1] == 16000):
self.count_stimuls += 1
elif (self.track[i][1] == 2):
self.count_click += 1
xy = self.__eval_max_peaks_fd_channel()
self.x_fd_peaks = xy[0]
self.y_fd_peaks = xy[1]
self.intervals = []
self.__eval_intervals()
self.file.close()
#%%
import numpy as np
import pandas as pd
import os
from HDF.HDFData import HDFArray
import math
# %%
# Берем первый файл данных
file = HDFArray('dataset/0.h5')
# Достаем таблицу трекера
table_track = file.getChannelTable('TRACK_MARKERS')
# Выделяем маркеры трекера
track_markers_columns = table_track[0]
# Выделяем массивы данных трекера
track_markers_data = table_track[1]
# %%
track_filter = []
for i in range(len(track_markers_data)):
if (track_markers_data[i][1] == 2):
track_filter.append(track_markers_data[i])
elif (i != len(track_markers_data) - 1 and track_markers_data[i][1] == 50
and (track_markers_data[i + 1][1] != 50)):
track_filter.append(track_markers_data[i])
elif (track_markers_data[i][1] == 0):
track_filter.append(track_markers_data[i])
elif (track_markers_data[i][1] == 16000):
track_filter.append(track_markers_data[i])
elif (track_markers_data[i][1] - track_markers_data[i][1] % 10 == 1000):
filteredPath = settings["paths"]["filtered"]
# опытным путем пришел к выводу, что данных номеров файлов нет среди данных
ban_files = [366, 269, 345, 344, 326, 325]
#for i in range(245, 368):
# if (i not in ban_files):
# print("id_files: ", i, "\nkeys: ", HDFArray("RawData/"+str(i) + '.h5').ListChannels().keys())
# %%
true_keys = ['EEG', 'TRACK_MARKERS', 'ECG', 'FD', 'PNVM']
all_files = 368 - 245 - len(ban_files)
good_files = 0
files = []
for i in range(245, 368):
if (i not in ban_files):
list_keys = HDFArray(path + str(i) + '.h5').ListChannels().keys()
k = 0
for key in true_keys:
if (key not in list_keys):
break
else:
k += 1
if (k == len(true_keys)):
files.append(str(i) + ".h5")
good_files += 1
print('file good of ', i, '.h5')
print('all files = ', all_files)
print("all good files = ", good_files)
# %%
import shutil
class FileAnalyzer:
def __init__(self, filename):
self.file = HDFArray(filename)
self.eeg = self.file.getChannelTable('EEG')[1]
self.ecg = self.file.getChannelTable('ECG')[1]
self.fd = self.file.getChannelTable('FD')[1]
self.pnvm = self.file.getChannelTable('PNVM')[1]
self.track = self.file.getChannelTable('TRACK_MARKERS')[1]
self.count_stimuls = 0
self.count_click = 0
for i in range(len(self.track)):
if (self.track[i][1] == 16000):
self.count_stimuls += 1
elif (self.track[i][1] == 2):
self.count_click += 1
xy = self.__eval_max_peaks_fd_channel()
self.x_fd_peaks = xy[0]
self.y_fd_peaks = xy[1]
self.intervals = []
self.__eval_intervals()
self.file.close()
def __eval_max_peaks_fd_channel(self):
count_diff_peaks = 2
_max, _min = pkd.peakdetect(self.fd, None, 50, 50)
xm = [p[0] for p in _max]
ym = [p[1] for p in _max]
if (abs(self.count_stimuls - len(xm)) > count_diff_peaks):
return xm, ym, 1
else:
diff = abs(len(xm) - self.count_stimuls)
if (len(xm) > self.count_stimuls):
xm = xm[:-diff]
ym = ym[:-diff]
elif (len(xm) < self.count_stimuls):
self.count_stimuls -= diff
return [xm, ym]
def get_max_peaks_fd_channel(self):
return [self.x_fd_peaks, self.y_fd_peaks]
def plot_fd_channel(self, a = None, b = None):
if (a is None or b is None):
plt.plot(self.fd)
plt.show()
else:
plt.plot(range(a,b), self.fd[a:b])
plt.show()
def get_count_synchrostimuls(self):
return self.count_stimuls
def get_count_click(self):
return self.count_click
def is_good_file(self):
if (self.count_stimuls != len(self.x_fd_peaks)):
return False
return True
def get_data_channels(self):
return [self.eeg, self.ecg, self.pnvm, self.fd]
def get_history(self):
for i in self.track:
if (i[1] == 16000):
print('start')
elif (i[1] - i[1]%10 == 1000):
print('pos_obj = ', i[2], i[3])
elif (i[1] == 2):
print('click')
elif (i[1] == 0):
print('clear and stop')
print('clear_obj = ', i[2], i[3])
def __get_interval(self, a, b):
obj_pos = [self.track[a][2],self.track[a][3]]
a = a + 1
int_eeg = self.eeg[a:b]
int_ecg = self.ecg[a:b]
int_pnvm = self.pnvm[a:b]
int_fd = self.fd[a:b]
int_track = self.track[a:b]
return IntervalExperiment(obj_pos,int_eeg, int_ecg,int_pnvm, int_fd, int_track)
def __eval_intervals(self):
count_stim = self.count_stimuls
stimuls = []
for i in range(len(self.track)):
if (self.track[i][1] == 16000):
stimuls.append(i)
count_stim -= 1
elif (self.track[i][1] == 2 and count_stim == 0):
stimuls.append(i)
for i in range(1, len(stimuls)-1):
interval = self.__get_interval(stimuls[i]-1, stimuls[i+1]-1)
self.intervals.append(interval)
def get_intervals(self):
return self.intervals
def get_min_length_intervals(self):
min_length_interval = self.intervals[0].get_length()
for interval in self.intervals:
if (interval.get_length() < min_length_interval):
min_length_interval = interval.get_length()
#print('min length interval = ', min_length_interval)
return min_length_interval
def trim_intervals_and_save(self,path, id_begin_interval, min_length_interval):
_id = id_begin_interval
y = []
for interval in self.intervals:
if (interval.is_good()):
interval.trim(0, min_length_interval)
filename = path + str(_id) + '.csv'
interval.save_to_csv(filename)
characteristics = interval.get_characteristics()
y.append(np.array([_id, characteristics[0], characteristics[1]]))
_id += 1
return np.array(y)