forked from ma-gilles/recovar
-
Notifications
You must be signed in to change notification settings - Fork 0
/
Copy pathassign_spike_datasets.py
187 lines (155 loc) · 8.74 KB
/
assign_spike_datasets.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
import recovar.config
from importlib import reload
import numpy as np
import matplotlib.pyplot as plt
import seaborn as sns
import plotly.graph_objs as go
import plotly.offline as py
from recovar.fourier_transform_utils import fourier_transform_utils
import jax.numpy as jnp
ftu = fourier_transform_utils(jnp)
from recovar import image_assignment, noise
from sklearn.metrics import confusion_matrix
from recovar import simulate_scattering_potential as ssp
from recovar import simulator, utils, image_assignment, noise, output, dataset
import prody
import pickle
import argparse
import sys
reload(simulator)
import scipy
import matplotlib.pyplot as plt
# added in
import cvxpy as cp
import seaborn as sns
sns.set_style("ticks")
sns.color_palette("colorblind")
def confusion_and_deconvolve(assignments, true_assignments, error_predicted):
# Compute the gamma from the note.
confus = confusion_matrix(assignments, true_assignments)
if confus.size > 1:
error_observed = (confus[1,0] + confus[0,1] ) / assignments.size
else:
error_observed = 0
# Apply deconvolution on the labels
observed_pop, deconvolve_pop, deconvolve_matrix = deconvolve_assignments(assignments, error_predicted)
return error_observed, observed_pop, deconvolve_pop, deconvolve_matrix
def deconvolve_assignments(assignments, error_predicted):
observed_pop = np.array([np.mean(assignments==0), np.mean(assignments==1)])
deconvolve_matrix = np.array( [ [1- error_predicted, error_predicted], [error_predicted, 1- error_predicted] ])
# Run constrained optimization
x = cp.Variable(2)
objective = cp.Minimize(cp.sum_squares(deconvolve_matrix @ x - observed_pop))
constraints = [0 <= x, x<= 1, sum(x) == 1]
prob = cp.Problem(objective, constraints)
prob.solve()
deconvolve_pop = x.value
return observed_pop, deconvolve_pop, deconvolve_matrix
def deconvolve_assignments_alt(assignments, error_predicted):
# Previous code used for this
observed_pop = np.array([np.mean(assignments==0), np.mean(assignments==1)])
deconvolve_matrix = np.array( [ [1- error_predicted, error_predicted], [error_predicted, 1- error_predicted] ])
deconvolve_pop = np.linalg.solve(deconvolve_matrix, observed_pop)
return observed_pop, deconvolve_pop, deconvolve_matrix
def main():
parser = argparse.ArgumentParser()
parser.add_argument("--output_folder", default="/mnt/home/levans/ceph/spike/recovar_experiments_10k_new", type=str)
args = parser.parse_args()
output_folder = args.output_folder
disc_type_infer = 'cubic'
grid_size = 256
# Load in things
file = open(output_folder + '/' + 'noise_levels.pkl','rb')
noise_levels = pickle.load(file)
error_observed = np.zeros(noise_levels.size)
error_predicted= np.zeros(noise_levels.size)
deconvolve_pop = np.zeros((noise_levels.size, 2))
deconvolve_pop_alt = np.zeros((noise_levels.size, 2))
observed_pop = np.zeros((noise_levels.size, 2))
observed_pop_soft = np.zeros((noise_levels.size, 2))
for idx, noise_level in enumerate(noise_levels):
dataset_folder = output_folder + '/' + f'dataset{idx}/'
print(f"Starting at noise level {idx} of {len(noise_levels)}")
# Load in simulation data
file = open(dataset_folder + '/' + 'sim_info.pkl','rb')
sim_info = pickle.load(file)
file.close()
# Load datasets and volumes
# Volumes are scaled so that images are normalized. So they have a slightly different scale for each dataset.
print(sim_info['volumes_path_root'])
volumes = simulator.load_volumes_from_folder(sim_info['volumes_path_root'], sim_info['grid_size'] , sim_info['trailing_zero_format_in_vol_name'], normalize=False )
gt_volumes = volumes * sim_info['scale_vol']
dataset_options = dataset.get_default_dataset_option()
dataset_options['particles_file'] = dataset_folder + f'particles.{grid_size}.mrcs'
dataset_options['ctf_file'] = dataset_folder + f'ctf.pkl'
dataset_options['poses_file'] = dataset_folder + f'poses.pkl'
cryo = dataset.load_dataset_from_dict(dataset_options, lazy = False)
# Compute likelihoods
batch_size = 100
image_cov_noise = np.asarray(noise.make_radial_noise(sim_info['noise_variance'], cryo.image_shape))
# transforming image assignment to log likelihoods
# NOTE: previous code I was using, for the weird plot, used this, without the 1/2
#log_likelihoods = -1*image_assignment.compute_image_assignment(cryo, gt_volumes, image_cov_noise, batch_size, disc_type = disc_type_infer).T
log_likelihoods = -0.5*image_assignment.compute_image_assignment(cryo, gt_volumes, image_cov_noise, batch_size, disc_type = disc_type_infer).T
# Compute hard assignments, hard assignment uncertainties
true_assignments = sim_info['image_assignment']
hard_assignments = jnp.argmax(log_likelihoods, axis = 1)
# Compute soft assignments and observed populations
log_likelihood_per_image = scipy.special.logsumexp(log_likelihoods, axis=1)
log_posteriors = log_likelihoods - log_likelihood_per_image.reshape(
log_likelihood_per_image.shape[0], 1
)
observed_pop_soft[idx, :] = np.exp(scipy.special.logsumexp(log_posteriors, axis=0))
observed_pop_soft[idx, :] /= observed_pop_soft[idx, :].sum()
error_predicted[idx] = image_assignment.estimate_false_positive_rate(cryo, gt_volumes, image_cov_noise, batch_size, disc_type = disc_type_infer)
error_observed[idx], observed_pop[idx, :], deconvolve_pop[idx, :], deconvolve_matrix = confusion_and_deconvolve(hard_assignments, true_assignments, error_predicted[idx])
_, deconvolve_pop_alt[idx], _ = deconvolve_assignments_alt(hard_assignments, error_predicted[idx])
print('o', error_observed[idx])
print('p', error_predicted[idx])
print('pops', deconvolve_pop[idx, :])
print('Observed pop:', observed_pop[idx])
print('Deconvolve mat:', deconvolve_matrix)
print('Deconvolved pop:', deconvolve_pop[idx, :])
# Dump results to file
likelihoods_assignments = { 'log_likelihoods': log_likelihoods,
'hard_assignments' : hard_assignments,
'true_assignments' : sim_info['image_assignment'],
}
recovar.utils.pickle_dump(likelihoods_assignments, dataset_folder + '/' + 'likelihoods_assignments.pkl')
recovar.utils.pickle_dump({'error_observed' : error_observed, \
'error_predicted' : error_predicted, \
'deconvolve_pop' : deconvolve_pop, \
'observed_pop_soft': observed_pop_soft, \
'observed_pop' : observed_pop, \
'deconvolve_pop_alt':deconvolve_pop_alt}, \
output_folder + '/' + 'pops_errors.pkl')
# Make a plot each time.
import matplotlib.pyplot as plt
plt.figure(figsize=(10, 6))
plt.semilogx(noise_levels[:idx+1], error_predicted[:idx+1], label='Analytical', color='blue', marker='o', markersize=6, linewidth=2)
plt.semilogx(noise_levels[:idx+1], error_observed[:idx+1], label='Observed', color='green', marker='s', markersize=6, linewidth=2)
plt.xlabel('Noise Level', fontsize=14)
plt.ylabel('False Positive Rate', fontsize=14)
plt.title('False Positive Rate vs. Noise Level', fontsize=16)
plt.legend(fontsize=12)
plt.grid(True, which='both', linestyle='--', linewidth=0.5)
plt.xticks(fontsize=12)
plt.yticks(fontsize=12)
plt.tight_layout()
plt.savefig(output_folder + '/' + 'curve.png')
plt.figure(figsize=(10, 6))
plt.semilogx(noise_levels[:idx+1], observed_pop[:idx+1, 0], label='Hard Assign', color='blue', marker='o', markersize=6, linewidth=2)
plt.semilogx(noise_levels[:idx+1], observed_pop_soft[:idx+1, 0], label='Soft Assign', color='orange', marker='o', markersize=6, linewidth=2)
plt.semilogx(noise_levels[:idx+1], deconvolve_pop[:idx+1, 0], label='Deconvolve', color='green', marker='s', markersize=6, linewidth=2)
plt.hlines(y=0.8, xmin=noise_levels[0], xmax=noise_levels[-1], label="True % Population", linestyle="--", color="k", linewidth=3.0)
plt.xlabel('Noise Level', fontsize=14)
plt.ylabel('% Population in state 1', fontsize=14)
plt.legend(fontsize=12)
plt.grid(True, which='both', linestyle='--', linewidth=0.5)
plt.xticks(fontsize=12)
plt.yticks(fontsize=12)
plt.tight_layout()
plt.savefig(output_folder + '/' + 'populations.png')
if __name__ == '__main__':
main()
print("Done")