-
Notifications
You must be signed in to change notification settings - Fork 1.3k
/
boxy.py
263 lines (237 loc) · 10.8 KB
/
boxy.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
257
258
259
260
261
262
263
# Authors: Kyle Mathewson, Jonathan Kuziek <kuziek@ualberta.ca>
#
# License: BSD-3-Clause
import re as re
import numpy as np
from ..base import BaseRaw
from ..meas_info import create_info
from ..utils import _mult_cal_one
from ...utils import logger, verbose, fill_doc, _check_fname
from ...annotations import Annotations
@fill_doc
def read_raw_boxy(fname, preload=False, verbose=None):
"""Reader for an optical imaging recording.
This function has been tested using the ISS Imagent I and II systems
and versions 0.40/0.84 of the BOXY recording software.
Parameters
----------
fname : str
Path to the BOXY data file.
%(preload)s
%(verbose)s
Returns
-------
raw : instance of RawBOXY
A Raw object containing BOXY data.
See Also
--------
mne.io.Raw : Documentation of attribute and methods.
"""
return RawBOXY(fname, preload, verbose)
@fill_doc
class RawBOXY(BaseRaw):
"""Raw object from a BOXY optical imaging file.
Parameters
----------
fname : str
Path to the BOXY data file.
%(preload)s
%(verbose)s
See Also
--------
mne.io.Raw : Documentation of attribute and methods.
"""
@verbose
def __init__(self, fname, preload=False, verbose=None):
logger.info('Loading %s' % fname)
# Read header file and grab some info.
start_line = np.inf
col_names = mrk_col = filetype = mrk_data = end_line = None
raw_extras = dict()
raw_extras['offsets'] = list() # keep track of our offsets
sfreq = None
fname = _check_fname(fname, 'read', True, 'fname')
with open(fname, 'r') as fid:
line_num = 0
i_line = fid.readline()
while i_line:
# most of our lines will be data lines, so check that first
if line_num >= start_line:
assert col_names is not None
assert filetype is not None
if '#DATA ENDS' in i_line:
# Data ends just before this.
end_line = line_num
break
if mrk_col is not None:
if filetype == 'non-parsed':
# Non-parsed files have different lines lengths.
crnt_line = i_line.rsplit(' ')[0]
temp_data = re.findall(
r'[-+]?\d*\.?\d+', crnt_line)
if len(temp_data) == len(col_names):
mrk_data.append(float(
re.findall(r'[-+]?\d*\.?\d+', crnt_line)
[mrk_col]))
else:
crnt_line = i_line.rsplit(' ')[0]
mrk_data.append(float(re.findall(
r'[-+]?\d*\.?\d+', crnt_line)[mrk_col]))
raw_extras['offsets'].append(fid.tell())
# now proceed with more standard header parsing
elif 'BOXY.EXE:' in i_line:
boxy_ver = re.findall(r'\d*\.\d+',
i_line.rsplit(' ')[-1])[0]
# Check that the BOXY version is supported
if boxy_ver not in ['0.40', '0.84']:
raise RuntimeError('MNE has not been tested with BOXY '
'version (%s)' % boxy_ver)
elif 'Detector Channels' in i_line:
raw_extras['detect_num'] = int(i_line.rsplit(' ')[0])
elif 'External MUX Channels' in i_line:
raw_extras['source_num'] = int(i_line.rsplit(' ')[0])
elif 'Update Rate (Hz)' in i_line or \
'Updata Rate (Hz)' in i_line:
# Version 0.40 of the BOXY recording software
# (and possibly other versions lower than 0.84) contains a
# typo in the raw data file where 'Update Rate' is spelled
# "Updata Rate. This will account for this typo.
sfreq = float(i_line.rsplit(' ')[0])
elif '#DATA BEGINS' in i_line:
# Data should start a couple lines later.
start_line = line_num + 3
elif line_num == start_line - 2:
# Grab names for each column of data.
raw_extras['col_names'] = col_names = re.findall(
r'\w+\-\w+|\w+\-\d+|\w+', i_line.rsplit(' ')[0])
if 'exmux' in col_names:
# Change filetype based on data organisation.
filetype = 'non-parsed'
else:
filetype = 'parsed'
if 'digaux' in col_names:
mrk_col = col_names.index('digaux')
mrk_data = list()
# raw_extras['offsets'].append(fid.tell())
elif line_num == start_line - 1:
raw_extras['offsets'].append(fid.tell())
line_num += 1
i_line = fid.readline()
assert sfreq is not None
raw_extras.update(
filetype=filetype, start_line=start_line, end_line=end_line)
# Label each channel in our data, for each data type (DC, AC, Ph).
# Data is organised by channels x timepoint, where the first
# 'source_num' rows correspond to the first detector, the next
# 'source_num' rows correspond to the second detector, and so on.
ch_names = list()
ch_types = list()
cals = list()
for det_num in range(raw_extras['detect_num']):
for src_num in range(raw_extras['source_num']):
for i_type, ch_type in [
('DC', 'fnirs_cw_amplitude'),
('AC', 'fnirs_fd_ac_amplitude'),
('Ph', 'fnirs_fd_phase')]:
ch_names.append(
f'S{src_num + 1}_D{det_num + 1} {i_type}')
ch_types.append(ch_type)
cals.append(np.pi / 180. if i_type == 'Ph' else 1.)
# Create info structure.
info = create_info(ch_names, sfreq, ch_types)
for ch, cal in zip(info['chs'], cals):
ch['cal'] = cal
# Determine how long our data is.
delta = end_line - start_line
assert len(raw_extras['offsets']) == delta + 1
if filetype == 'non-parsed':
delta //= (raw_extras['source_num'])
super(RawBOXY, self).__init__(
info, preload, filenames=[fname], first_samps=[0],
last_samps=[delta - 1], raw_extras=[raw_extras], verbose=verbose)
# Now let's grab our markers, if they are present.
if mrk_data is not None:
mrk_data = np.array(mrk_data, float)
# We only want the first instance of each trigger.
prev_mrk = 0
mrk_idx = list()
duration = list()
tmp_dur = 0
for i_num, i_mrk in enumerate(mrk_data):
if i_mrk != 0 and i_mrk != prev_mrk:
mrk_idx.append(i_num)
if i_mrk != 0 and i_mrk == prev_mrk:
tmp_dur += 1
if i_mrk == 0 and i_mrk != prev_mrk:
duration.append((tmp_dur + 1) / sfreq)
tmp_dur = 0
prev_mrk = i_mrk
onset = np.array(mrk_idx) / sfreq
description = mrk_data[mrk_idx]
annot = Annotations(onset, duration, description)
self.set_annotations(annot)
def _read_segment_file(self, data, idx, fi, start, stop, cals, mult):
"""Read a segment of data from a file.
Boxy file organises data in two ways, parsed or un-parsed.
Regardless of type, output has (n_montages x n_sources x n_detectors
+ n_marker_channels) rows, and (n_timepoints x n_blocks) columns.
"""
source_num = self._raw_extras[fi]['source_num']
detect_num = self._raw_extras[fi]['detect_num']
start_line = self._raw_extras[fi]['start_line']
end_line = self._raw_extras[fi]['end_line']
filetype = self._raw_extras[fi]['filetype']
col_names = self._raw_extras[fi]['col_names']
offsets = self._raw_extras[fi]['offsets']
boxy_file = self._filenames[fi]
# Non-parsed multiplexes sources, so we need source_num times as many
# lines in that case
if filetype == 'parsed':
start_read = start_line + start
stop_read = start_read + (stop - start)
else:
assert filetype == 'non-parsed'
start_read = start_line + start * source_num
stop_read = start_read + (stop - start) * source_num
assert start_read >= start_line
assert stop_read <= end_line
# Possible detector names.
detectors = 'ABCDEFGHIJKLMNOPQRSTUVWXYZ'[:detect_num]
# Loop through our data.
one = np.zeros((len(col_names), stop_read - start_read))
with open(boxy_file, 'r') as fid:
# Just a more efficient version of this:
# ii = 0
# for line_num, i_line in enumerate(fid):
# if line_num >= start_read:
# if line_num >= stop_read:
# break
# # Grab actual data.
# i_data = i_line.strip().split()
# one[:len(i_data), ii] = i_data
# ii += 1
fid.seek(offsets[start_read - start_line], 0)
for oo in one.T:
i_data = fid.readline().strip().split()
oo[:len(i_data)] = i_data
# in theory we could index in the loop above, but it's painfully slow,
# so let's just take a hopefully minor memory hit
if filetype == 'non-parsed':
ch_idxs = [col_names.index(f'{det}-{i_type}')
for det in detectors
for i_type in ['DC', 'AC', 'Ph']]
one = one[ch_idxs].reshape( # each "time point" multiplexes srcs
len(detectors), 3, -1, source_num
).transpose( # reorganize into (det, source, DC/AC/Ph, t) order
0, 3, 1, 2
).reshape( # reshape the way we store it (det x source x DAP, t)
len(detectors) * source_num * 3, -1)
else:
assert filetype == 'parsed'
ch_idxs = [col_names.index(f'{det}-{i_type}{si + 1}')
for det in detectors
for si in range(source_num)
for i_type in ['DC', 'AC', 'Ph']]
one = one[ch_idxs]
# Place our data into the data object in place.
_mult_cal_one(data, one, idx, cals, mult)