-
Notifications
You must be signed in to change notification settings - Fork 0
/
edf_convert.py
256 lines (196 loc) · 9.36 KB
/
edf_convert.py
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
149
150
151
152
153
154
155
156
157
158
159
160
161
162
163
164
165
166
167
168
169
170
171
172
173
174
175
176
177
178
179
180
181
182
183
184
185
186
187
188
189
190
191
192
193
194
195
196
197
198
199
200
201
202
203
204
205
206
207
208
209
210
211
212
213
214
215
216
217
218
219
220
221
222
223
224
225
226
227
228
229
230
231
232
233
234
235
236
237
238
239
240
241
242
243
244
245
246
247
248
249
250
251
252
253
254
255
256
# -*- coding: utf-8 -*-
### --------------- IMPORTS --------------- ###
import os, sys, json
import tables
import pyedflib
import numpy as np
from scipy import signal
from tqdm import tqdm
### --------------------------------------- ###
class EdfConvert:
""" Class for conversion of .edf files to .h5 format.
"""
def __init__(self, prop_dict):
"""
Parameters
----------
prop_dict : Dict, with properties from config.json file
Returns
-------
None.
"""
# get values from dictionary
for key, value in prop_dict.items():
setattr(self, key, value)
self.winsize = int(self.new_fs*self.win)
self.filelist = list(filter(lambda k: '.edf' in k, os.listdir(self.main_path)))
self.h5_filelist = list(filter(lambda k: '.h5' in k, os.listdir(self.main_path)))
def _read_edf(self, file_name):
"""
Opens edf file using pyedflib and returns a reference object.
----------------------------------------------------------------------
Parameters
----------
file_name : Str, file name
Returns
-------
fread : pyedflib obj
"""
fread = pyedflib.EdfReader(os.path.join(self.main_path, file_name))
return fread
def edf_check(self, file_name, read_length=1000):
"""
Read small parts of an edf file. Read samples from
start, mid and end across all channels of the edf file.
----------------------------------------------------------------------
Parameters
----------
file_name : Str, file name
read_length: Int, Number of samples to be read for each segment
Optional, Default = 1000
Returns
-------
"""
# open edf file and check if sampling rate is higher than user set sampling rate (default 100Hz)
f_edf = self._read_edf(file_name)
sampling_rate = f_edf.getSampleFrequencies()
if np.any(sampling_rate < self.new_fs):
raise Exception(f'--> Original sampling rate {sampling_rate} is lower than new sampling rate ({self.new_fs} Hz).\n')
if self.selected_channel_idx is None:
# check whether selected channels exist
channel_labels = f_edf.getSignalLabels()
if not set(self.selected_channels) <= set(channel_labels):
raise Exception(f'--> Selected channels {self.selected_channels} were not found in File: {file_name}. Got {channel_labels} instead.\n')
# get channel index
ch_idx = []
for channel_name in self.selected_channels:
ch_idx.append(channel_labels.index(channel_name))
else:
ch_idx = self.selected_channel_idx
if not set(ch_idx) <= set(np.arange(len(sampling_rate))):
raise Exception(f'--> Selected channels were not found in file: {file_name}. ' \
f'Available channels are {np.arange(len(sampling_rate))} but got channel index: {ch_idx} instead.\n')
# read signal samples from start, mid, and end portions
for i in ch_idx:
signal_length = f_edf.getNSamples()[i]
f_edf.readSignal(chn=i, start=0, n=read_length)
f_edf.readSignal(chn=i, start=int(signal_length/2), n=read_length)
f_edf.readSignal(chn=i, start=int(signal_length - read_length - 1) , n=read_length)
del f_edf
def edf_to_h5(self, file_name):
"""
Convert an edf to h5 file.
h5 file shape:
1st-dimension, X = nSamples/Y
2nd-dimension, Y = win * new_fs
3rd-dimension, Z = number of channels
Where 'nSamples' is the number of samples in one channel of the edf file.
----------------------------------------------------------------------
Parameters
----------
file_name : Str, file name
Returns
-------
"""
# open edf reader and get number of rows
# assuming recorded time is the same across channels different sampling rate should not affect number of rows
f_edf = self._read_edf(file_name)
sampling_rate = f_edf.getSampleFrequencies()
down_factor = int(sampling_rate[0]/self.new_fs)
nrows = int(f_edf.getNSamples()[0]/down_factor/self.winsize)
# get channel index
if self.selected_channel_idx is None:
channel_labels = f_edf.getSignalLabels()
ch_idx = []
for channel_name in self.selected_channels:
ch_idx.append(channel_labels.index(channel_name))
else:
ch_idx = self.selected_channel_idx
# open tables object for saving
with tables.open_file(os.path.join(self.main_path, file_name.replace('.edf','.h5')), mode='w') as fsave:
# create data store
ds = fsave.create_earray(fsave.root, 'data', tables.Float64Atom(), shape=[nrows, self.winsize, 0])
# iterate over channels to preprocess signal (decimate, scale, reshape) and append to datastore
for ch_num in ch_idx: # iterate over channels
fs = f_edf.getSampleFrequency(ch_num)
down_factor = int(fs/self.new_fs)
data = signal.decimate(f_edf.readSignal(ch_num), down_factor) * self.scale
# trim data to winsize*rows and reshape
data = np.reshape(data[:self.winsize*nrows], (-1, self.winsize, 1))
ds.append(data)
del f_edf
def all_files(self, func):
"""
Run func operation on all edf files in parent edf directory.
----------------------------------------------------------------------
Parameters
----------
func : Function or method for manipulation of one edf file
Returns
-------
bool : False/True for successful/unsuccessful operation
"""
# get file list and iterate over all files for conversion
try:
for file in tqdm((self.filelist), desc='Progress', file=sys.stdout, total=len(self.filelist)):
func(file)
return False
except Exception as err:
print(f'\n -> Error! File: {file} could not be read.\n {str(err)}\n')
return True
def run_conversion(config_path='config.json'):
# load properties from configuration file
openpath = open( config_path, 'r').read();
prop_dict = json.loads(openpath)
# get parent directory
main_path = input('Please enter path of folder containing edf files: \n')
if not os.path.isdir(main_path):
print('---> Path:', "'" + main_path + "'", 'is not valid.\n Please enter a valid path.')
sys.exit()
else:
prop_dict.update({'main_path':main_path})
# init object
obj = EdfConvert(prop_dict)
if len(obj.filelist) == 0:
print('--> No Edf files were detected.')
sys.exit()
# Verify how to proceed
if len(obj.h5_filelist) > 0:
options = ['y', 'n']
answer = ''
while answer not in options:
answer = input(f'H5 files were detected and proceeding might overwrite them. Continue? {str(options)}\n')
if answer not in options:
print('\n---> Input error: Please choose one of the following options:', str(options) +'.',
'This was received instead:', str(answer)+'\n')
if answer == 'n':
print('\n---> No Further Action Will Be Performed.\n')
sys.exit()
print('\n---------------------------------------------------------------------')
print(f'{len(obj.filelist)} files were detected. Initiate Error Check: \n')
success = obj.all_files(obj.edf_check)
print('\n------------------------ Error Check Finished -----------------------')
print('---------------------------------------------------------------------\n')
if success == False:
print('-> File Check Completed Successfully.\n')
else:
print('--> Warning! File check was not successful. Please remove/edit problematic edf files or adjust the config file.\n')
sys.exit()
# Verify how to proceed
options = ['y', 'n']
answer = ''
while answer not in options:
answer = input(f'Would you like to proceed with File Conversion? {options}\n')
if answer not in options:
print('\n---> Input error: Please choose one of the following options:', str(options) +'.',
'This was received instead:', str(answer)+'\n')
if answer == 'n':
print('\n---> No Further Action Will Be Performed.\n')
sys.exit()
elif answer == 'y':
print('\n-------------------------------------------------------------------------------')
print('------------------------ Initiating edf -> h5 Conversion ----------------------\n')
obj.all_files(obj.edf_to_h5)
print('\n************************* Conversion Completed *************************\n')
if __name__ == '__main__':
run_conversion(config_path='config.json')