An easy-to-use platform for EEG experimentation in the classroom
rewriting the python dependency within brainwaves
the objective is to clean up and organize the python calls.
- pyimport contains all the python modules to be used
- utils is for all the function definitions
- cells are independent commands to be made (maybe rename to commands)
pipes.js and function.js were responsible for managing the state of the python kernel. they won't be needed in this new paradigm
+281
+90
app/utils/pyodide/commands.py
+90
app/utils/pyodide/commands.py
···
1
+
import * as path from 'path';
2
+
import { readFileSync } from 'fs';
3
+
4
+
// -----------------------------
5
+
// Imports and Utility functions
6
+
7
+
export const imports = () =>
8
+
readFileSync(path.join(__dirname, '/utils/pyodide/pyimport.py'), 'utf8');
9
+
10
+
export const utils = () =>
11
+
readFileSync(path.join(__dirname, '/utils/jupyter/utils.py'), 'utf8');
12
+
13
+
export const loadCSV = (filePathArray: Array<string>) =>
14
+
[
15
+
`files = [${filePathArray.map(filePath => formatFilePath(filePath))}]`,
16
+
`replace_ch_names = None`,
17
+
`raw = load_data(files, replace_ch_names)`
18
+
].join('\n');
19
+
20
+
// ---------------------------
21
+
// MNE-Related Data Processing
22
+
export const loadCleanedEpochs = (filePathArray: Array<string>) =>
23
+
[
24
+
`files = [${filePathArray.map(filePath => formatFilePath(filePath))}]`,
25
+
`clean_epochs = concatenate_epochs([read_epochs(file) for file in files])`,
26
+
`conditions = OrderedDict({key: [value] for (key, value) in clean_epochs.event_id.items()})`
27
+
].join('\n');
28
+
29
+
// NOTE: this command includes a ';' to prevent returning data
30
+
export const filterIIR = (lowCutoff: number, highCutoff: number) =>
31
+
`raw.filter(${lowCutoff}, ${highCutoff}, method='iir');`;
32
+
33
+
export const epochEvents = (
34
+
eventIDs: { [string]: number },
35
+
tmin: number,
36
+
tmax: number,
37
+
reject?: Array<string> | string = 'None'
38
+
) => {
39
+
const command = [
40
+
`event_id = ${JSON.stringify(eventIDs)}`,
41
+
`tmin=${tmin}`,
42
+
`tmax=${tmax}`,
43
+
`baseline= (tmin, tmax)`,
44
+
`picks = None`,
45
+
`reject = ${reject}`,
46
+
'events = find_events(raw)',
47
+
`raw_epochs = Epochs(raw, events=events, event_id=event_id,
48
+
tmin=tmin, tmax=tmax, baseline=baseline, reject=reject, preload=True,
49
+
verbose=False, picks=picks)`,
50
+
`conditions = OrderedDict({key: [value] for (key, value) in raw_epochs.event_id.items()})`
51
+
].join('\n');
52
+
return command;
53
+
};
54
+
55
+
export const requestEpochsInfo = (variableName: string) =>
56
+
`get_epochs_info(${variableName})`;
57
+
58
+
export const requestChannelInfo = () =>
59
+
`[ch for ch in clean_epochs.ch_names if ch != 'Marker']`;
60
+
61
+
// -----------------------------
62
+
// Plot functions
63
+
64
+
export const cleanEpochsPlot = () =>
65
+
`raw_epochs.plot(scalings='auto', n_epochs=6, title="Clean Data", events=None)`;
66
+
67
+
export const plotPSD = () => `raw.plot_psd(fmin=1, fmax=30)`;
68
+
69
+
export const plotTopoMap = () => `plot_topo(clean_epochs, conditions)`;
70
+
71
+
export const plotERP = (channelIndex: number) =>
72
+
`X, y = plot_conditions(clean_epochs, ch_ind=${channelIndex}, conditions=conditions,
73
+
ci=97.5, n_boot=1000, title='', diff_waveform=None)`;
74
+
75
+
export const saveEpochs = (workspaceDir: string, subject: string) =>
76
+
`raw_epochs.save(${formatFilePath(
77
+
path.join(
78
+
workspaceDir,
79
+
'Data',
80
+
subject,
81
+
'EEG',
82
+
`${subject}-cleaned-epo.fif`
83
+
)
84
+
)})`;
85
+
86
+
// -------------------------------------------
87
+
// Helper methods
88
+
89
+
const formatFilePath = (filePath: string) =>
90
+
`"${filePath.replace(/\\/g, '/')}"`;
+16
app/utils/pyodide/pyimport.py
+16
app/utils/pyodide/pyimport.py
···
1
+
from time import time, strftime, gmtime
2
+
import os
3
+
from collections import OrderedDict
4
+
from glob import glob
5
+
6
+
import numpy as np
7
+
import pandas as pd # maybe we can remove this dependency
8
+
import seaborn as sns
9
+
from matplotlib import pyplot as plt
10
+
11
+
from mne import (Epochs, RawArray, concatenate_raws, concatenate_epochs,
12
+
create_info, find_events, read_epochs, set_eeg_reference)
13
+
from mne.channels import read_montage
14
+
15
+
16
+
plt.style.use(fivethirtyeight)
+175
app/utils/pyodide/utils.py
+175
app/utils/pyodide/utils.py
···
1
+
from glob import glob
2
+
import os
3
+
from collections import OrderedDict
4
+
from mne import create_info, concatenate_raws, viz
5
+
from mne.io import RawArray
6
+
from mne.channels import read_montage
7
+
import pandas as pd
8
+
import numpy as np
9
+
import seaborn as sns
10
+
from matplotlib import pyplot as plt
11
+
12
+
sns.set_context('talk')
13
+
sns.set_style('white')
14
+
15
+
16
+
def load_data(fnames, sfreq=128., replace_ch_names=None):
17
+
"""Load CSV files from the /data directory into a Raw object.
18
+
19
+
Args:
20
+
fnames (array): CSV filepaths from which to load data
21
+
22
+
Keyword Args:
23
+
sfreq (float): EEG sampling frequency
24
+
replace_ch_names (dict or None): dictionary containing a mapping to
25
+
rename channels. Useful when an external electrode was used.
26
+
27
+
Returns:
28
+
(mne.io.array.array.RawArray): loaded EEG
29
+
"""
30
+
31
+
raw = []
32
+
print(fnames)
33
+
for fname in fnames:
34
+
# read the file
35
+
data = pd.read_csv(fname, index_col=0)
36
+
37
+
data = data.dropna()
38
+
39
+
# get estimation of sampling rate and use to determine sfreq
40
+
# yes, this could probably be improved
41
+
srate = 1000 / (data.index.values[1] - data.index.values[0])
42
+
if srate >= 200:
43
+
sfreq = 256
44
+
else:
45
+
sfreq = 128
46
+
47
+
# name of each channel
48
+
ch_names = list(data.columns)
49
+
50
+
# indices of each channel
51
+
ch_ind = list(range(len(ch_names)))
52
+
53
+
if replace_ch_names is not None:
54
+
ch_names = [c if c not in replace_ch_names.keys()
55
+
else replace_ch_names[c] for c in ch_names]
56
+
57
+
# type of each channels
58
+
ch_types = ['eeg'] * (len(ch_ind) - 1) + ['stim']
59
+
montage = read_montage('standard_1005')
60
+
61
+
# get data and exclude Aux channel
62
+
data = data.values[:, ch_ind].T
63
+
64
+
# create MNE object
65
+
info = create_info(ch_names=ch_names, ch_types=ch_types,
66
+
sfreq=sfreq, montage=montage)
67
+
raw.append(RawArray(data=data, info=info))
68
+
69
+
# concatenate all raw objects
70
+
raws = concatenate_raws(raw)
71
+
72
+
return raws
73
+
74
+
75
+
def plot_topo(epochs, conditions=OrderedDict()):
76
+
palette = sns.color_palette("hls", len(conditions) + 1)
77
+
evokeds = [epochs[name].average() for name in (conditions)]
78
+
79
+
evoked_topo = viz.plot_evoked_topo(
80
+
evokeds, vline=None, color=palette[0:len(conditions)], show=False)
81
+
evoked_topo.patch.set_alpha(0)
82
+
evoked_topo.set_size_inches(10, 8)
83
+
for axis in evoked_topo.axes:
84
+
for line in axis.lines:
85
+
line.set_linewidth(2)
86
+
87
+
legend_loc = 0
88
+
labels = [e.comment if e.comment else 'Unknown' for e in evokeds]
89
+
legend = plt.legend(labels, loc=legend_loc, prop={'size': 20})
90
+
txts = legend.get_texts()
91
+
for txt, col in zip(txts, palette):
92
+
txt.set_color(col)
93
+
94
+
return evoked_topo
95
+
96
+
97
+
def plot_conditions(epochs, ch_ind=0, conditions=OrderedDict(), ci=97.5, n_boot=1000,
98
+
title='', palette=None,
99
+
diff_waveform=(4, 3)):
100
+
"""Plot Averaged Epochs with ERP conditions.
101
+
102
+
Args:
103
+
epochs (mne.epochs): EEG epochs
104
+
105
+
Keyword Args:
106
+
conditions (OrderedDict): dictionary that contains the names of the
107
+
conditions to plot as keys, and the list of corresponding marker
108
+
numbers as value. E.g.,
109
+
110
+
conditions = {'Non-target': [0, 1],
111
+
'Target': [2, 3, 4]}
112
+
113
+
ch_ind (int): index of channel to plot data from
114
+
ci (float): confidence interval in range [0, 100]
115
+
n_boot (int): number of bootstrap samples
116
+
title (str): title of the figure
117
+
palette (list): color palette to use for conditions
118
+
ylim (tuple): (ymin, ymax)
119
+
diff_waveform (tuple or None): tuple of ints indicating which
120
+
conditions to subtract for producing the difference waveform.
121
+
If None, do not plot a difference waveform
122
+
123
+
Returns:
124
+
(matplotlib.figure.Figure): figure object
125
+
(list of matplotlib.axes._subplots.AxesSubplot): list of axes
126
+
"""
127
+
if isinstance(conditions, dict):
128
+
conditions = OrderedDict(conditions)
129
+
130
+
if palette is None:
131
+
palette = sns.color_palette("hls", len(conditions) + 1)
132
+
133
+
X = epochs.get_data()
134
+
times = epochs.times
135
+
y = pd.Series(epochs.events[:, -1])
136
+
fig, ax = plt.subplots()
137
+
138
+
for cond, color in zip(conditions.values(), palette):
139
+
sns.tsplot(X[y.isin(cond), ch_ind], time=times, color=color,
140
+
n_boot=n_boot, ci=ci)
141
+
142
+
if diff_waveform:
143
+
diff = (np.nanmean(X[y == diff_waveform[1], ch_ind], axis=0) -
144
+
np.nanmean(X[y == diff_waveform[0], ch_ind], axis=0))
145
+
ax.plot(times, diff, color='k', lw=1)
146
+
147
+
ax.set_title(epochs.ch_names[ch_ind])
148
+
ax.axvline(x=0, color='k', lw=1, label='_nolegend_')
149
+
150
+
ax.set_xlabel('Time (s)')
151
+
ax.set_ylabel('Amplitude (uV)')
152
+
ax.set_xlabel('Time (s)')
153
+
ax.set_ylabel('Amplitude (uV)')
154
+
155
+
# Round y axis tick labels to 2 decimal places
156
+
# ax.yaxis.set_major_formatter(FormatStrFormatter('%.2f'))
157
+
158
+
if diff_waveform:
159
+
legend = (['{} - {}'.format(diff_waveform[1], diff_waveform[0])] +
160
+
list(conditions.keys()))
161
+
else:
162
+
legend = conditions.keys()
163
+
ax.legend(legend)
164
+
sns.despine()
165
+
plt.tight_layout()
166
+
167
+
if title:
168
+
fig.suptitle(title, fontsize=20)
169
+
170
+
fig.set_size_inches(10, 8)
171
+
172
+
return fig, ax
173
+
174
+
def get_epochs_info(epochs):
175
+
return [*[{x: len(epochs[x])} for x in epochs.event_id], {"Drop Percentage": round((1 - len(epochs.events)/len(epochs.drop_log)) * 100, 2)}, {"Total Epochs": len(epochs.events)}]