-
Notifications
You must be signed in to change notification settings - Fork 8
/
plotter.py
253 lines (190 loc) · 9.99 KB
/
plotter.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
import matplotlib.pyplot as plt
import numpy as np
import torch
import pdb
import os
from utils import batchify_random_sample
class Plotter(object):
def __init__(self, time, truth=None, base_fontsize=14):
self.dt = np.diff(time)[0]
self.time = time
self.fontsize={'ticklabel' : base_fontsize-2,
'label' : base_fontsize,
'title' : base_fontsize+2,
'suptitle' : base_fontsize+4}
self.colors = {'linc_red' : '#E84924',
'linc_blue' : '#37A1D0'}
self.truth = truth
#------------------------------------------------------------------------------
def plot_summary(self, model, dl, num_average=200, ix=None, mode='traces', save_dir=None):
'''
plot_summary(data, truth=None, num_average=100, ix=None)
Plot summary figures for dataset and ground truth if available. Create a batch
from one sample by repeating a certain number of times, and average across them.
Arguments:
- data (torch.Tensor) : dataset
- truth (dict) : ground truth dictionary
- num_average (int) : number of samples from posterior to average over
- ix (int) : index of data samples to make summary plot from
Returns:
- fig_dict : dict of summary figures
'''
plt.close()
figs_dict = {}
data = dl.dataset.tensors[0]
batch_example, ix = batchify_random_sample(data=data, batch_size=num_average, ix=ix)
batch_example = batch_example.to(model.device)
figs_dict['ix'] = ix
model.eval()
with torch.no_grad():
recon, (factors, inputs) = model(batch_example)
orig = batch_example[0].cpu().numpy()
# print(batch_example.shape, data.shape, recon['data'].shape)
# pdb.set_trace()
if mode=='traces':
figs_dict['traces'] = self.plot_traces(recon['data'].mean(dim=0).detach().cpu().numpy(), orig, mode='activity', norm=True)
figs_dict['traces'].suptitle('Actual fluorescence trace vs.\nestimated mean for a sampled trial')
elif mode=='video':
# TODO
# figs_dict['videos'] = self.plot_video(recon['data'].mean(dim=0).detach().cpu().numpy(), orig)
save_video_dir = save_dir + 'videos/'
if not os.path.exists(save_video_dir):
os.mkdir(save_video_dir)
self.plot_video(recon['data'].mean(dim=0).detach().cpu().numpy(), orig, save_folder = save_video_dir)
if self.truth:
if 'rates' in self.truth.keys():
recon_rates = recon['rates'].mean(dim=1).cpu().numpy()
true_rates = self.truth['rates'][ix]
figs_dict['truth_rates'] = self.plot_traces(recon_rates, true_rates, mode='rand')
figs_dict['truth_rates'].suptitle('Reconstructed vs ground-truth rate function')
if 'latent' in self.truth.keys():
pred_factors = factors.mean(dim=1).cpu().numpy()
true_factors = self.truth['latent'][ix]
# pdb.set_trace()
figs_dict['truth_factors'] = self.plot_traces(pred_factors, true_factors, num_traces=true_factors.shape[-1], ncols=1)
figs_dict['truth_factors'].suptitle('Reconstructed vs ground-truth factors')
else:
figs_dict['factors'] = self.plot_factors(factors.mean(dim=1).cpu().numpy())
if 'spikes' in self.truth.keys():
if 'spikes' in recon.keys():
recon_spikes = recon['spikes'].mean(dim=1).cpu().numpy()
true_spikes = self.truth['spikes'][ix]
figs_dict['truth_spikes'] = self.plot_traces(recon_spikes, true_spikes, mode='rand')
figs_dict['truth_spikes'].suptitle('Reconstructed vs ground-truth rate function')
else:
figs_dict['factors'] = self.plot_factors(factors.mean(dim=1).cpu().numpy())
if inputs is not None:
figs_dict['inputs'] = self.plot_inputs(inputs.mean(dim=1).cpu().numpy())
return figs_dict
#------------------------------------------------------------------------------W
#------------------------------------------------------------------------------
def plot_traces(self, pred, true, figsize=(8,8), num_traces=12, ncols=2, mode=None, norm=False, pred_logvar=None):
'''
Plot trace and compare to ground truth
Arguments:
- pred (np.array): array of predicted values to plot (dims: num_steps x num_cells)
- true (np.array) : array of true values to plot (dims: num_steps x num_cells)
- figsize (2-tuple) : figure size (width, height) in inches (default = (8, 8))
- num_traces (int) : number of traces to plot (default = 24)
- ncols (int) : number of columns in figure (default = 2)
- mode (string) : mode to select subset of traces. Options: 'activity', 'rand', None.
'Activity' plots the the num_traces/2 most active traces and num_traces/2
least active traces defined sorted by mean value in trace
- norm (bool) : normalize predicted and actual values (default=True)
- pred_logvar (np.array) : array of predicted values log-variance (dims: num_steps x num_cells) (default= None)
'''
num_cells = pred.shape[-1]
nrows = int(num_traces/ncols)
fig, axs = plt.subplots(figsize=figsize, nrows=nrows, ncols=ncols, sharex=True, sharey=True)
axs = np.ravel(axs)
if mode == 'rand':
idxs = np.random.choice(list(range(num_cells)), size=num_traces, replace=False)
idxs.sort()
elif mode == 'activity':
idxs = true.max(axis=0).argsort()[-num_traces:]
else:
idxs = list(range(num_cells))
for ii, (ax,idx) in enumerate(zip(axs,idxs)):
if norm is True:
true_norm= (true[:, idx] - np.mean(true[:, idx]))/np.std(true[:, idx])
pred_norm= (pred[:, idx] - np.mean(pred[:, idx]))/np.std(pred[:, idx])
plt.sca(ax)
plt.plot(self.time, true_norm, lw=2, color=self.colors['linc_red'])
plt.plot(self.time, pred_norm, lw=2, color=self.colors['linc_blue'])
else:
plt.sca(ax)
plt.plot(self.time, true[:, idx], lw=2, color=self.colors['linc_red'])
plt.plot(self.time, pred[:, idx], lw=2, color=self.colors['linc_blue'])
fig.subplots_adjust(wspace=0.1, hspace=0.1)
plt.legend(['Actual', 'Reconstructed'])
return fig
#------------------------------------------------------------------------------
def plot_video(self, pred, true, save_folder): #
# TODO
# pass
num_frames = true.shape[1]
num_frames_pred = pred.shape[1]
for t in range(num_frames):
fig, (ax1, ax2) = plt.subplots(1, 2, sharey=True)
neg1 = ax1.imshow(pred[0,t,:,:])
neg2 = ax2.imshow(true[0,t,:,:])
neg1.set_clim(vmin=0, vmax=2)
neg2.set_clim(vmin=0, vmax=2)
fig.savefig(save_folder+str(t)+'.png')
plt.close(fig)
#------------------------------------------------------------------------------
def plot_factors(self, factors, max_in_col=5, figsize=(8,8)):
'''
plot_factors(max_in_col=5, figsize=(8,8))
Plot inferred factors in a grid
Arguments:
- max_in_col (int) : maximum number of subplots in a column
- figsize (tuple of 2 ints) : figure size in inches
Returns
- figure
'''
steps_size, factors_size = factors.shape
nrows = min(max_in_col, factors_size)
ncols = int(np.ceil(factors_size/max_in_col))
fig, axs = plt.subplots(nrows=nrows, ncols=ncols, figsize=figsize)
axs = np.ravel(axs)
fmin = factors.min()
fmax = factors.max()
for jx in range(factors_size):
plt.sca(axs[jx])
plt.plot(self.time, factors[:, jx])
plt.ylim(fmin-0.1, fmax+0.1)
if jx%ncols == 0:
plt.ylabel('Activity')
else:
plt.ylabel('')
axs[jx].set_yticklabels([])
if (jx - jx%ncols)/ncols == (nrows-1):
plt.xlabel('Time (s)')
else:
plt.xlabel('')
axs[jx].set_xticklabels([])
fig.suptitle('Factors 1-%i for a sampled trial.'%factors.shape[1])
fig.subplots_adjust(wspace=0.1, hspace=0.1)
return fig
#------------------------------------------------------------------------------
def plot_inputs(self, inputs, fig_width=8, fig_height=1.5):
'''
plot_inputs(fig_width=8, fig_height=1.5)
Plot inferred inputs
Arguments:
- fig_width (int) : figure width in inches
- fig_height (int) : figure height in inches
'''
steps_size, inputs_size = inputs.shape
figsize = (fig_width, fig_height*inputs_size)
fig, axs = plt.subplots(nrows=inputs_size, figsize=figsize)
fig.suptitle('Input to the generator for a sampled trial', y=1.2)
for jx in range(inputs_size):
if inputs_size > 1:
plt.sca(axs[jx])
else:
plt.sca(axs)
plt.plot(self.time, inputs[:, jx])
plt.xlabel('time (s)')
return fig