-
Notifications
You must be signed in to change notification settings - Fork 0
/
utils_plot.py
338 lines (275 loc) · 11.5 KB
/
utils_plot.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
264
265
266
267
268
269
270
271
272
273
274
275
276
277
278
279
280
281
282
283
284
285
286
287
288
289
290
291
292
293
294
295
296
297
298
299
300
301
302
303
304
305
306
307
308
309
310
311
312
313
314
315
316
317
318
319
320
321
322
323
324
325
326
327
328
329
330
331
332
333
334
335
336
337
338
"""
"""
# %%
from pathlib import Path
import numpy as np
import matplotlib.pyplot as plt
import pickle
import mne
from config import RESULTS_DIR, get_paths
from utils_csc import get_subject_dipole
# from dripp.trunc_norm_kernel.model import TruncNormKernel
# def plot_csc(cdl_model, raw_csc, allZ,
# plot_acti_histo=False, shift_acti=True,
# activation_tstart=0, df_dripp=None,
# save_dir=Path('.'), title=None, show=True):
# """Plot the returns of CSC model.
# Parameters
# ----------
# cdl_model : instance of alphacsc.ConvolutionalDictionaryLearning
# raw_csc : mne.io.Raw
# The raw data on which CDL was run.
# allZ :
# plot_acti_histo : bool
# if True, plot the histogram of activations
# shift_acti : bool
# if True, roll to put activation to the peak amplitude time in the atom
# activation_tstart : float
# XXX I don't like you hard code 1.7 here.
# default is 0
# df_dripp : pandas.DataFrame
# possible DriPP results
# default is None
# save_dir : instance of pathlib.Path
# path to saving directory
# title : str
# title to put on figure
# show : bool
# Show figures at the end or not.
# Returns
# -------
# figs : list
# The list of generated matplotlib figures.
# """
# fontsize = 12
# n_atoms_per_fig = 5
# n_plot_per_atom = 3 + plot_acti_histo + (df_dripp is not None)
# n_atoms_est = allZ.shape[1]
# info = raw_csc.info
# sfreq = raw_csc.info['sfreq']
# atom_duration = cdl_model.v_hat_.shape[-1] / raw_csc.info['sfreq']
# figsize = (15, 10)
# atoms_in_figs = np.arange(0, n_atoms_est + 1, n_atoms_per_fig)
# atoms_in_figs = list(zip(atoms_in_figs[:-1], atoms_in_figs[1:]))
# figs = []
# for fig_idx, (atoms_start, atoms_stop) in enumerate(atoms_in_figs, start=1):
# fig, axes = plt.subplots(
# n_plot_per_atom, n_atoms_per_fig, figsize=figsize)
# figs.append(fig)
# fig.suptitle(title, fontsize=fontsize)
# for i_atom, kk in enumerate(range(atoms_start, atoms_stop)):
# ax = axes[0, i_atom]
# ax.set_title("Atom #" + str(kk), fontsize=fontsize)
# # Spatial pattern
# u_hat = cdl_model.u_hat_[kk]
# mne.viz.plot_topomap(u_hat, info, axes=ax, show=False)
# if i_atom == 0:
# ax.set_ylabel("Spatial", labelpad=86, fontsize=fontsize)
# # Temporal pattern
# ax = axes[1, i_atom]
# v_hat = cdl_model.v_hat_[kk]
# t = np.arange(v_hat.size) / sfreq
# ax.plot(t, v_hat)
# ax.grid(True)
# ax.set_xlim(0, atom_duration) # crop x axis
# if i_atom == 0:
# ax.set_ylabel("Temporal", labelpad=14, fontsize=fontsize)
# # Power Spectral Density (PSD)
# ax = axes[2, i_atom]
# psd = np.abs(np.fft.rfft(v_hat, n=256)) ** 2
# frequencies = np.linspace(0, sfreq / 2.0, len(psd))
# ax.semilogy(frequencies, psd, label="PSD", color="k")
# ax.set_xlim(0, 40) # crop x axis
# ax.set_xlabel("Frequencies (Hz)", fontsize=fontsize)
# ax.grid(True)
# if i_atom == 0:
# ax.set_ylabel("Power Spectral Density", labelpad=13,
# fontsize=fontsize)
# if plot_acti_histo:
# # Atom's activations
# ax = axes[3, i_atom]
# z_hat = allZ[:, kk, :]
# if shift_acti:
# # roll to put activation to the peak amplitude time
# shift = np.argmax(np.abs(v_hat))
# z_hat = np.roll(z_hat, shift, axis=1)
# z_hat[:, :shift] = 0 # pad with 0
# t1 = np.arange(allZ.shape[2]) / sfreq - activation_tstart
# ax.plot(t1, z_hat.T)
# ax.set_xlabel("Time (s)", fontsize=fontsize)
# if i_atom == 0:
# ax.set_ylabel("Atom's activations",
# labelpad=7, fontsize=fontsize)
# if (df_dripp is not None):
# # Atom's learned intensity
# ax = axes[4, i_atom]
# # get DriPP results
# columns = ['baseline', 'alpha', 'm', 'sigma', 'lower', 'upper']
# res = df_dripp[columns][df_dripp['atom'] == kk].iloc[0]
# baseline = res['baseline']
# lower, upper = res['lower'], res['upper']
# if np.isnan(baseline):
# ax.text(0.5, 0.5, "no activation",
# horizontalalignment='center',
# verticalalignment='center',
# transform=ax.transAxes)
# else:
# # xx = np.linspace(-0.5, upper, 500)
# xx = t1
# try:
# # XXX
# # for subject CC110033, atom 15, DriPP returns [nan]
# # for alpha and kernel shape, but 0 for baseline
# # define kernel function
# alpha = res['alpha'][0]
# m, sigma = res['m'][0], res['sigma'][0]
# kernel = TruncNormKernel(lower, upper, m, sigma)
# yy = baseline + alpha * kernel.eval(xx)
# except AssertionError:
# yy = baseline * np.ones(xx.shape)
# # plot learned intensity
# ax.plot(xx, yy, label='button')
# # ax.set_xlim(-0.5, upper)
# if i_atom == 0:
# # intensity_ax = ax
# ax.set_ylabel("Intensity", labelpad=7, fontsize=fontsize)
# # else:
# # # rescale inteisty axe
# # intensity_ax.get_shared_y_axes().join(intensity_ax, ax)
# # ax.autoscale()
# ax.legend(fontsize=fontsize, handlelength=1)
# fig.tight_layout()
# fig.subplots_adjust(top=0.88)
# fig_name = f"atoms_part_{fig_idx}.pdf"
# fig.savefig(save_dir / fig_name, dpi=300)
# fig.savefig(save_dir / (fig_name.replace(".pdf", ".png")),
# dpi=300)
# if show:
# plt.show()
# return figs
# # %%
def plot_atoms_single_sub(atom_df, subject_id, sfreq=150., plot_psd=False, plot_dipole=False, save_dir=None):
"""Plot the atoms of a single subject.
Parameters
----------
atom_df : pandas.DataFrame
each row is an atom, has minimum columns 'subject_id', 'atom_id', 'u_hat', 'v_hat'
subject_id : str
Returns
-------
"""
df = atom_df[atom_df['subject_id'] == subject_id].reset_index()
n_atoms = df['atom_id'].nunique()
# get info
file_name = RESULTS_DIR / subject_id / 'CSCraw_0.5s_20atoms.pkl'
cdl_model, info, _, _ = pickle.load(open(file_name, "rb"))
meg_indices = mne.pick_types(info, meg='grad')
info = mne.pick_info(info, meg_indices)
if plot_dipole:
# get dipole
dip = get_subject_dipole(subject_id, cdl_model=cdl_model, info=info)
epochFif, transFif, bemFif = get_paths(subject_id)
# shape of the final figure
fontsize = 12
n_columns = min(5, n_atoms)
split = int(np.ceil(n_atoms / n_columns))
n_plots = 2 + plot_psd + plot_dipole
figsize = (4 * n_columns, 3 * n_plots * split)
fig, axes = plt.subplots(n_plots * split, n_columns, figsize=figsize)
axes = np.atleast_2d(axes)
for ii, row in df.iterrows():
kk = row.atom_id
# Select the axes to display the current atom
i_row, i_col = ii // n_columns, ii % n_columns
it_axes = iter(axes[i_row * n_plots:(i_row + 1) * n_plots, i_col])
ax = next(it_axes)
ax.set_title('Atom % d' % kk, fontsize=fontsize, pad=0)
# Plot the spatial map of the atom using mne topomap
mne.viz.plot_topomap(data=row.u_hat, pos=info, axes=ax, show=False)
if i_col == 0:
ax.set_ylabel('Spatial', labelpad=30, fontsize=fontsize)
# Plot the temporal pattern of the atom
v_hat = row.v_hat
ax = next(it_axes)
ax.plot(np.arange(v_hat.shape[0]) / sfreq, v_hat)
atom_duration = v_hat.shape[-1] / sfreq
ax.set_xlim(0, atom_duration)
if i_col == 0:
temporal_ax = ax
ax.set_ylabel('Temporal', fontsize=fontsize)
if i_col > 0:
ax.get_yaxis().set_visible(False)
temporal_ax.get_shared_y_axes().join(temporal_ax, ax)
ax.autoscale()
if plot_psd:
ax = next(it_axes)
psd = np.abs(np.fft.rfft(v_hat, n=256)) ** 2
frequencies = np.linspace(0, sfreq / 2.0, len(psd))
ax.semilogy(frequencies, psd, label="PSD", color="k")
ax.set_xlim(0, 40) # crop x axis
ax.set_xlabel("Frequencies (Hz)")
ax.grid(True)
if i_col == 0:
ax.set_ylabel("Power Spectral Density", labelpad=8)
if plot_dipole:
ax = next(it_axes)
dip.plot_locations(str(transFif), '01', subjects_dir,
idx=kk, ax=ax, show_all=False)
pass
fig.tight_layout()
if save_dir is not None:
plt.savefig(save_dir + f'/atoms_subject_{subject_id}.jpg')
plt.show()
return fig
def plot_mean_atom(df, info, sfreq=150., plot_psd=False, plot_acti_histo=False, plot_dipole=False):
"""
"""
n_atoms = len(set(df['label'].values))
n_columns = n_atoms
n_plots = 2 + plot_psd + plot_acti_histo + plot_dipole
figsize = (4 * n_columns, 3 * n_plots)
fig, axes = plt.subplots(n_plots, n_columns, figsize=figsize)
axes = np.atleast_2d(axes)
for ii, row in df.iterrows():
label, u_hat, v_hat, z_hat = row.label, row.u_hat, row.v_hat, row.z_hat
# Select the axes to display the current atom
i_row, i_col = ii // n_columns, ii % n_columns
it_axes = iter(axes[i_row * n_plots:(i_row + 1) * n_plots, i_col])
ax = next(it_axes)
ax.set_title(f'Class label {label}', pad=0)
# Plot the spatial map of the atom using mne topomap
mne.viz.plot_topomap(data=u_hat, pos=info, axes=ax, show=False)
if i_col == 0:
ax.set_ylabel('Spatial', labelpad=30)
# Plot the temporal pattern of the atom
ax = next(it_axes)
ax.plot(np.arange(v_hat.shape[0]) / sfreq, v_hat)
atom_duration = v_hat.shape[-1] / 150.
ax.set_xlim(0, atom_duration)
if i_col == 0:
temporal_ax = ax
ax.set_ylabel('Temporal')
if i_col > 0:
ax.get_yaxis().set_visible(False)
temporal_ax.get_shared_y_axes().join(temporal_ax, ax)
ax.autoscale()
if plot_psd:
ax = next(it_axes)
psd = np.abs(np.fft.rfft(v_hat, n=256)) ** 2
frequencies = np.linspace(0, sfreq / 2.0, len(psd))
ax.semilogy(frequencies, psd, label="PSD", color="k")
ax.set_xlim(0, 40) # crop x axis
ax.set_xlabel("Frequencies (Hz)")
ax.grid(True)
if i_col == 0:
ax.set_ylabel("Power Spectral Density", labelpad=8)
if plot_acti_histo:
# XXX
pass
if plot_dipole:
# XXX
pass
fig.tight_layout()
plt.show()
return fig
# %%