Repository navigation
Expand file tree
/
Copy pathbenchmark_optimization_evaluation.py
More file actions
128 lines (111 loc) · 5.61 KB
/
Copy pathbenchmark_optimization_evaluation.py
File metadata and controls
128 lines (111 loc) · 5.61 KB
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
import model_builder
import sys, os, math
import pickle
import matplotlib
from matplotlib.pyplot import tight_layout
matplotlib.rcParams['ps.useafm'] = True
matplotlib.rcParams['pdf.use14corefonts'] = True
matplotlib.rcParams['text.usetex'] = True
import matplotlib.pyplot as plt
import matplotlib.patches as mpatches
from matplotlib.collections import PatchCollection
import matplotlib_tools
def main():
input_name = 'output/optimized_cell_positions_benchmark_'
output_name = 'output/optimizer/'
os.makedirs(output_name, exist_ok=True)
datasets = []
# List of body files, list of synapse files, right eye (False) or left eye (True)
datasets.append((['../raw_connectome/FIB-25/annotations-body.json'], ['../raw_connectome/FIB-25/annotations-synapse.json'], False))
datasets.append((['../raw_connectome/FIB-19/body.txt'], ['../raw_connectome/FIB-19/export-fib19.json'], False))
# Parse
dataset_bodies, dataset_synapses = model_builder.parse_datasets_to_model(datasets)
# [dataset][cell_type][(index, id, (y,x))]
orig_dataset_known_positions = []
# [dataset][cell_type][(index, id)]
orig_dataset_unknown_positions = []
for i in range(0, len(dataset_bodies)):
index = 0
sub_known_positions = dict()
sub_unknown_positions = dict()
bodies = dataset_bodies[i]
for body_key in list(bodies.keys()):
body = bodies[body_key]
if not body[0] in sub_known_positions:
sub_known_positions[body[0]] = []
if not body[0] in sub_unknown_positions:
sub_unknown_positions[body[0]] = []
if body[1] == None:
sub_unknown_positions[body[0]].append((index, body_key))
else:
sub_known_positions[body[0]].append((index, body_key, body[1]))
index = index + 1
orig_dataset_known_positions.append(sub_known_positions)
orig_dataset_unknown_positions.append(sub_unknown_positions)
results = []
for i in range(0,11):
dataset_known_positions, _ = pickle.load(open(input_name+str(i)+'.pickle', 'rb'))
results.append(dataset_known_positions)
cell_types = set()
for known_positions in results[0]:
for cell_type in list(known_positions.keys()):
cell_types.add(cell_type)
count_correct = [[0 for j in range(0, 1+2*(len(results[0])+1))] for i in range(0, len(results))]
total_position_count = 0
known_position_count = 0
dataset_known_position_count = [0 for i in range(0, len(dataset_bodies))]
dataset_total_position_count = [0 for i in range(0, len(dataset_bodies))]
for i in range(0, len(dataset_bodies)):
bodies = dataset_bodies[i]
for body_key in list(bodies.keys()):
body = bodies[body_key]
if not body[1] == None:
known_position_count += 1
dataset_known_position_count[i] += 1
total_position_count += 1
dataset_total_position_count[i] += 1
for cell_type in cell_types:
for i in range(0, len(results[0])):
if cell_type in results[0][i].keys():
reference_known_positions = results[0][i][cell_type]
for reference_known_position in reference_known_positions:
count_correct[0][1] += 1
count_correct[0][2+i] += 1
for j in range(1,11):
for compare_known_position in results[j][i][cell_type]:
if (compare_known_position[1] == reference_known_position[1] and
compare_known_position[3] == reference_known_position[3]):
count_correct[j][1] += 1
count_correct[j][2+i] += 1
print('Here 1' + str(compare_known_position) + str(reference_known_position))
for cell_type in cell_types:
for i in range(0, len(orig_dataset_known_positions)):
if cell_type in orig_dataset_known_positions[i].keys():
reference_known_positions = orig_dataset_known_positions[i][cell_type]
for reference_known_position in reference_known_positions:
for j in range(0,11):
for compare_known_position in results[j][i][cell_type]:
if (compare_known_position[1] == reference_known_position[1] and
compare_known_position[3] == reference_known_position[2]):
count_correct[j][len(orig_dataset_known_positions)+2] += 1
count_correct[j][len(orig_dataset_known_positions)+3+i] += 1
xlabels = [i*10 for i in range(0,11)]
for j in reversed(range(0,11)):
count_correct[j][0] = (known_position_count-(known_position_count/10)*j)#/total_position_count * 100.0
print(count_correct)
plt.figure(figsize=(6, 6))
plt.plot(xlabels, count_correct)
ax = plt.axes()
plt.legend(['Reference positions', 'Combined', 'FIB-25', 'FIB-19', 'Combined (annotated)', 'FIB-25 (annotated)', 'FIB-19 (annotated)'])
plt.xlabel('Known positions removed [\%]')
plt.ylabel('Matching positions')
plt.xlim([0,100])
plt.ylim([0,2000])
ax.spines['right'].set_visible(False)
ax.spines['top'].set_visible(False)
plt.savefig(output_name+'/optimization_benchmark.pdf', tight_layout=True)
with open(output_name+'/optimization_benchmark.txt', 'w') as txt_file:
txt_file.write(str(count_correct)+'\n')
plt.show()
if __name__ == "__main__":
main()