-
Notifications
You must be signed in to change notification settings - Fork 1
/
scenario_PLOS_table_V4_10m.py
372 lines (333 loc) · 18.6 KB
/
scenario_PLOS_table_V4_10m.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
339
340
341
342
343
344
345
346
347
348
349
350
351
352
353
354
355
356
357
358
359
360
361
362
363
364
365
366
367
368
369
370
371
372
import sys
import numpy as np
import time
import pickle
import matplotlib.pyplot as plt
from CoronaTestingSimulation import Corona_Simulation
from Statistics import Corona_Simulation_Statistics
import multiprocessing
'''
Scenario 1
Test all individuals of a population
'''
# whether to print plotdata
PRINTPLOTDATA = True
# default plot font sizes
SMALL_SIZE = 12
MEDIUM_SIZE = 14
BIGGER_SIZE = 16
plt.rc('font', size=SMALL_SIZE) # controls default text sizes
plt.rc('axes', titlesize=SMALL_SIZE) # fontsize of the axes title
plt.rc('axes', labelsize=MEDIUM_SIZE) # fontsize of the x and y labels
plt.rc('xtick', labelsize=SMALL_SIZE) # fontsize of the tick labels
plt.rc('ytick', labelsize=SMALL_SIZE) # fontsize of the tick labels
plt.rc('legend', fontsize=SMALL_SIZE) # legend fontsize
plt.rc('figure', titlesize=BIGGER_SIZE) # fontsize of the figure title
# name for the data dump and plots
def getName(scale_factor_pop, scale_factor_test, success_rate_test=0.99):
name = 'PLOS_V4_10m_scenario1_scalepop{}_scaletest{}'.format(scale_factor_pop, scale_factor_test)
if success_rate_test != 0.99:
name += '_{}'.format(success_rate_test)
return name
def worker(return_dict, sample_size, prob_sick, success_rate_test, false_posivite_rate, test_strategy,
num_simultaneous_tests, test_duration, group_size, scale_factor_pop,
tests_repetitions, test_result_decision_strategy, number_of_instances, country):
'''
worker function for multiprocessing
performs the same test tests_repetitions many times and returns expected values and standard deviations
'''
stat_test = Corona_Simulation_Statistics(prob_sick, success_rate_test,
false_posivite_rate, test_strategy,
test_duration, group_size,
tests_repetitions, test_result_decision_strategy,
scale_factor_pop)
stat_test.statistical_analysis(sample_size, num_simultaneous_tests, number_of_instances)
print('Calculated {} for {} prob sick {}'.format(test_strategy, country, prob_sick))
print('scaled to {} population and {} simulataneous tests\n'.format(sample_size, num_simultaneous_tests))
# gather results
worker_dict = {}
worker_dict['e_num_tests'] = stat_test.e_number_of_tests*scale_factor_pop
worker_dict['e_time'] = stat_test.e_time*scale_factor_pop
worker_dict['e_num_confirmed_sick_individuals'] = stat_test.e_num_confirmed_sick_individuals*scale_factor_pop
worker_dict['e_false_positive_rate'] = stat_test.e_false_positive_rate
worker_dict['e_ratio_of_sick_found'] = stat_test.e_ratio_of_sick_found
worker_dict['e_num_confirmed_per_test'] = stat_test.e_num_confirmed_per_test
worker_dict['e_num_sent_to_quarantine'] = stat_test.e_num_sent_to_quarantine
worker_dict['sd_num_tests'] = stat_test.sd_number_of_tests*scale_factor_pop
worker_dict['sd_time'] = stat_test.sd_time*scale_factor_pop
worker_dict['sd_false_positive_rate'] = stat_test.sd_false_positive_rate
worker_dict['sd_ratio_of_sick_found'] = stat_test.sd_ratio_of_sick_found
worker_dict['sd_num_confirmed_per_test'] = stat_test.sd_num_confirmed_per_test
worker_dict['sd_num_sent_to_quarantine'] = stat_test.sd_num_sent_to_quarantine
return_dict['{}_{}_{}'.format(test_strategy, country, prob_sick)] = worker_dict
def calculation():
start = time.time()
randomseed = 19
np.random.seed(randomseed)
probabilities_sick = [0.01]
success_rate_test = 0.99
false_posivite_rate = 0.01
tests_repetitions = 1
test_result_decision_strategy = 'max'
number_of_instances = 20
test_duration = 5
# optimal group sizes in order individual, two level, binary splitting, RBS, purim, sobel
optimal_group_sizes = {}
if success_rate_test == 0.99:
optimal_group_sizes[0.001] = [1, 32, 32, 32, 32, 32]
optimal_group_sizes[0.0025] = [1, 23, 32, 32, 32, 32]
optimal_group_sizes[0.005] = [1, 16, 32, 32, 32, 32]
optimal_group_sizes[0.0075] = [1, 12, 32, 32, 32, 32]
optimal_group_sizes[0.01] = [1, 10, 32, 32, 27, 31]
optimal_group_sizes[0.025] = [1, 7, 16, 30, 14, 30]
optimal_group_sizes[0.05] = [1, 5, 8, 15, 10, 27]
optimal_group_sizes[0.1] = [1, 4, 4, 8, 7, 20]
optimal_group_sizes[0.15] = [1, 3, 4, 6, 6, 32]
optimal_group_sizes[0.2] = [1, 3, 2, 1, 5, 30]
optimal_group_sizes[0.25] = [1, 3, 2, 1, 5, 28]
optimal_group_sizes[0.3] = [1, 3, 1, 1, 1, 19]
optimal_group_sizes[0.5] = [1, 3, 1, 1, 1, 19]
elif success_rate_test == 0.75:
optimal_group_sizes[0.001] = [1, 32, 32, 32, 32, 32]
optimal_group_sizes[0.0025] = [1, 21, 32, 32, 32, 32]
optimal_group_sizes[0.005] = [1, 18, 32, 32, 32, 32]
optimal_group_sizes[0.0075] = [1, 15, 32, 32, 32, 32]
optimal_group_sizes[0.01] = [1, 12, 32, 32, 31, 32]
optimal_group_sizes[0.025] = [1, 8, 32, 32, 18, 30]
optimal_group_sizes[0.05] = [1, 6, 32, 32, 12, 32]
optimal_group_sizes[0.1] = [1, 5, 32, 31, 8, 8]
optimal_group_sizes[0.15] = [1, 4, 32, 32, 7, 6]
optimal_group_sizes[0.2] = [1, 4, 32, 32, 32, 4]
optimal_group_sizes[0.25] = [1, 4, 32, 32, 32, 3]
optimal_group_sizes[0.3] = [1, 30, 32, 32, 32, 32]
# strings identifiying the test strategies
test_strategies = [
'individual-testing',
'two-stage-testing',
'binary-splitting',
'RBS',
'purim',
'sobel'
]
# use scale_factor_pop = 10 for the original results in the paper
# use scale_factor_pop = 100 for much faster calculation and little loss of accuracy
countries = {}
countries['DE'] = {'population': 10000000, 'tests_per_day': 100000,
'scale_factor_pop': 1, 'scale_factor_test': 1}
num_countries = len(countries.keys())
print('ref values for individual testing:')
for country in countries:
print('{} {}'.format(country, int(countries[country]['population'] / countries[country]['tests_per_day'])))
print('\n')
manager = multiprocessing.Manager()
return_dict = manager.dict()
e_num_tests = np.zeros((len(test_strategies), num_countries, len(probabilities_sick)))
e_time = np.zeros((len(test_strategies), num_countries, len(probabilities_sick)))
e_false_positive_rate = np.zeros((len(test_strategies), num_countries, len(probabilities_sick)))
e_num_confirmed_sick_individuals = np.zeros((len(test_strategies), num_countries, len(probabilities_sick)))
e_ratio_of_sick_found = np.zeros((len(test_strategies), num_countries, len(probabilities_sick)))
e_num_confirmed_per_test = np.zeros((len(test_strategies), num_countries, len(probabilities_sick)))
e_num_sent_to_quarantine = np.zeros((len(test_strategies), num_countries, len(probabilities_sick)))
sd_num_tests = np.zeros((len(test_strategies), num_countries, len(probabilities_sick)))
sd_time = np.zeros((len(test_strategies), num_countries, len(probabilities_sick)))
sd_false_positive_rate = np.zeros((len(test_strategies), num_countries, len(probabilities_sick)))
sd_ratio_of_sick_found = np.zeros((len(test_strategies), num_countries, len(probabilities_sick)))
sd_num_confirmed_per_test = np.zeros((len(test_strategies), num_countries, len(probabilities_sick)))
sd_num_sent_to_quarantine = np.zeros((len(test_strategies), num_countries, len(probabilities_sick)))
jobs = []
for i, test_strategy in enumerate(test_strategies):
for j, country in enumerate(countries.keys()):
for k, prob_sick in enumerate(probabilities_sick):
group_size = optimal_group_sizes[prob_sick][i]
scale_factor_pop = countries[country]['scale_factor_pop']
scale_factor_test = countries[country]['scale_factor_test']
sample_size = int(countries[country]['population'] / scale_factor_pop / scale_factor_test)
num_simultaneous_tests = int(
np.ceil(countries[country]['tests_per_day']/scale_factor_test*test_duration/24.0))
p = multiprocessing.Process(target=worker, args=(return_dict, sample_size, prob_sick,
success_rate_test, false_posivite_rate, test_strategy, num_simultaneous_tests,
test_duration, group_size, scale_factor_pop, tests_repetitions, test_result_decision_strategy,
number_of_instances, country))
jobs.append(p)
p.start()
for proc in jobs:
proc.join()
# gather results
for i, test_strategy in enumerate(test_strategies):
for j, country in enumerate(countries.keys()):
for k, prob_sick in enumerate(probabilities_sick):
worker_dict = return_dict['{}_{}_{}'.format(test_strategy, country, prob_sick)]
e_num_tests[i, j, k] = worker_dict['e_num_tests']
e_time[i, j, k] = worker_dict['e_time']
e_num_confirmed_sick_individuals[i, j, k] = worker_dict['e_num_confirmed_sick_individuals']
e_false_positive_rate[i, j, k] = worker_dict['e_false_positive_rate']
e_ratio_of_sick_found[i, j, k] = worker_dict['e_ratio_of_sick_found']
e_num_confirmed_per_test[i, j, k] = worker_dict['e_num_confirmed_per_test']
e_num_sent_to_quarantine[i, j, k] = worker_dict['e_num_sent_to_quarantine']
sd_num_tests[i, j, k] = worker_dict['sd_num_tests']
sd_time[i, j, k] = worker_dict['sd_time']
sd_false_positive_rate[i, j, k] = worker_dict['sd_false_positive_rate']
sd_ratio_of_sick_found[i, j, k] = worker_dict['sd_ratio_of_sick_found']
sd_num_confirmed_per_test[i, j, k] = worker_dict['sd_num_confirmed_per_test']
sd_num_sent_to_quarantine[i, j, k] = worker_dict['sd_num_sent_to_quarantine']
sample_sizes = [countries[country]['population'] for country in countries.keys()]
daily_tests_per_1m = [countries[country]['tests_per_day']/countries[country]
['population']*1000000 for country in countries.keys()]
print('daily_test_per_1m {}'.format(daily_tests_per_1m))
runtime = time.time()-start
print('calculating took {}s'.format(runtime))
# save data to allow plotting without doing the whole calculation again.
data = {
'randomseed': randomseed,
'probabilities_sick': probabilities_sick,
'success_rate_test ': success_rate_test,
'false_posivite_rate': false_posivite_rate,
'tests_repetitions': tests_repetitions,
'test_result_decision_strategy': test_result_decision_strategy,
'test_strategies': test_strategies,
'countries': countries,
'number_of_instances': number_of_instances,
'test_duration': test_duration,
'group_size': group_size,
'e_num_tests ': e_num_tests,
'e_time': e_time,
'e_false_positive_rate': e_false_positive_rate,
'e_num_confirmed_sick_individuals': e_num_confirmed_sick_individuals,
'e_ratio_of_sick_found': e_ratio_of_sick_found,
'e_num_confirmed_per_test': e_num_confirmed_per_test,
'e_num_sent_to_quarantine': e_num_sent_to_quarantine,
'sd_num_tests': sd_num_tests,
'sd_time': sd_time,
'sd_false_positive_rate': sd_false_positive_rate,
'sd_ratio_of_sick_found': sd_ratio_of_sick_found,
'sd_num_confirmed_per_test': sd_num_confirmed_per_test,
'sd_num_sent_to_quarantine': sd_num_sent_to_quarantine,
'sample_sizes': sample_sizes,
'daily_tests_per_1m': daily_tests_per_1m,
'runtime': runtime,
}
filename = getName(countries['DE']['scale_factor_pop'],
countries['DE']['scale_factor_test'],
success_rate_test)
path = 'data/{}.pkl'.format(filename)
with open(path, 'wb+') as fp:
pickle.dump(data, fp)
print('saved data as {}'.format(path))
return filename
def plotting(filename, prob_sick_plot_index, saveFig=0):
# load data
datapath = 'data/{}.pkl'.format(filename)
with open(datapath, 'rb') as fp:
data = pickle.load(fp)
figpath = 'plots/{}'.format(filename)
# extract relevant parameters from data
test_strategies = data['test_strategies']
daily_tests_per_1m = data['daily_tests_per_1m']
countries = data['countries']
probabilities_sick = data['probabilities_sick']
e_time = data['e_time']
sd_time = data['sd_time']
e_num_confirmed_per_test = data['e_num_confirmed_per_test']
sd_num_confirmed_per_test = data['sd_num_confirmed_per_test']
# plotting
markers = ['o', '*', '^', '+', 's', 'd', 'v', '<', '>']
colors = ['C0', 'C1', 'C2', 'C3', 'C4', 'C5', 'C6', 'C7', 'C8', 'C9']
linestyles = ['-', '-', '-', '-', '-', '--']
labels = ['Individual testing', '2-level pooling',
'Binary splitting', 'Recursive binary splitting', 'Purim', 'Sobel-R1']
######## prob sick / sick persons per test ########
plt.figure()
for i, test_strategy in enumerate(test_strategies):
for j in [0]: # it's the same for all countries
plt.plot(probabilities_sick, e_num_confirmed_per_test[i, j, :],
label=labels[i], marker=markers[i], color=colors[i], linestyle=linestyles[i])
plt.errorbar(probabilities_sick, e_num_confirmed_per_test[i, j, :],
yerr=sd_num_confirmed_per_test[i, j, :], ecolor='k', linestyle='None', capsize=5)
plt.xlabel('infection rate')
plt.ylabel('exp. number of identified cases per test')
plt.xticks([0.001, 0.01, 0.05, 0.1, 0.15, 0.2, 0.25, 0.3], [
'0.1% ', ' 1%', '5%', '10%', '15%', '20%', '25%', '30%', ])
plt.legend(loc='lower right', fontsize=11)
if PRINTPLOTDATA:
print(figpath+'psi{}_probsick_ppt.pdf'.format(prob_sick_plot_index))
print("%20s" % "probabilities_sick", "".join(map(lambda x: "%7.4f " % x, probabilities_sick)))
for i, test_strategy in enumerate(test_strategies):
print("%20s" % test_strategy, "".join(map(lambda x: "%7.5f " % x, e_num_confirmed_per_test[i, 0, :])))
if saveFig:
plt.savefig(figpath+'psi{}_probsick_ppt.pdf'.format(prob_sick_plot_index), bbox_inches='tight')
######## prob sick / sick persons per test (Zoomed)########
plt.figure()
for i, test_strategy in enumerate(test_strategies):
for j in [0]: # it's the same for all countries
plt.plot(probabilities_sick[:7], e_num_confirmed_per_test[i, j, :][:7],
label=labels[i], marker=markers[i], color=colors[i], linestyle=linestyles[i])
plt.errorbar(probabilities_sick[:7], e_num_confirmed_per_test[i, j, :][:7],
yerr=sd_num_confirmed_per_test[i, j, :][:7], ecolor='k', linestyle='None', capsize=5)
plt.xlabel('infection rate')
plt.ylabel('exp. number of identified cases per test')
plt.xticks([0.001, 0.01, 0.025, 0.05], ['0.1%', '1%', '2.5%', '5%'])
if PRINTPLOTDATA:
print(figpath+'psi{}_probsick_ppt.pdf'.format(prob_sick_plot_index))
print(probabilities_sick)
for i, test_strategy in enumerate(test_strategies):
print(test_strategy, e_num_confirmed_per_test[i, 0, :])
if saveFig:
plt.savefig(figpath+'psi{}_probsick_ppt_zoomed.pdf'.format(prob_sick_plot_index), bbox_inches='tight')
######## test per 1M / expected time to test all ########
plt.figure()
for i, test_strategy in enumerate(test_strategies):
plt.plot(daily_tests_per_1m, e_time[i, :, prob_sick_plot_index],
label=labels[i], marker=markers[i], color=colors[i], linestyle=linestyles[i])
plt.errorbar(daily_tests_per_1m, e_time[i, :, prob_sick_plot_index],
yerr=sd_time[i, :, prob_sick_plot_index], ecolor='k',
linestyle='None', capsize=5)
plt.xticks(daily_tests_per_1m, ['{} {}'.format(country, int(daily_tests_per_1m[i]))
for i, country in enumerate(countries)], rotation=55)
plt.ylabel('exp. time to test population [days]')
plt.xlabel('daily tests / 1M population.')
plt.legend(loc='upper right')
if PRINTPLOTDATA:
print(figpath+'psi{}_testsper1M_time.pdf'.format(prob_sick_plot_index))
print("infection rate: %7.4f" % probabilities_sick[prob_sick_plot_index])
print(" "*20, "".join(map(lambda x: "%7s " % x, countries)))
print("%20s" % "daily_test_per_1m", "".join(map(lambda x: "%7.2f " % x, daily_tests_per_1m)))
for i, test_strategy in enumerate(test_strategies):
print("%20s" % test_strategy, "".join(map(lambda x: "%7.2f " % x, e_time[i, :, prob_sick_plot_index])))
if saveFig:
plt.savefig(figpath+'psi{}_testsper1M_time.pdf'.format(prob_sick_plot_index), bbox_inches='tight')
######## test per 1M / expected time to test all (Zoomed)########
plt.figure()
for i, test_strategy in enumerate(test_strategies):
plt.plot(daily_tests_per_1m, e_time[i, :, prob_sick_plot_index],
label=labels[i], marker=markers[i], color=colors[i], linestyle=linestyles[i])
plt.errorbar(daily_tests_per_1m, e_time[i, :, prob_sick_plot_index],
yerr=sd_time[i, :, prob_sick_plot_index], ecolor='k',
linestyle='None', capsize=5)
plt.xticks(daily_tests_per_1m, ['{} {}'.format(country, int(daily_tests_per_1m[i]))
for i, country in enumerate(countries)], rotation=55)
plt.ylabel('exp. time to test population [days]')
plt.xlabel('daily tests / 1M population.')
if PRINTPLOTDATA:
print(figpath+'psi{}_testsper1M_time.pdf'.format(prob_sick_plot_index))
print(daily_tests_per_1m)
for i, test_strategy in enumerate(test_strategies):
print(test_strategy, e_time[i, :, prob_sick_plot_index])
plt.ylim([0, 1250])
if saveFig:
plt.savefig(figpath+'psi{}_testsper1M_time_zoomed.pdf'.format(prob_sick_plot_index), bbox_inches='tight')
if __name__ == "__main__":
recalculate = True
if recalculate:
# either do calculations
filename = calculation()
else:
# or use precalculated data
scale_factor_pop = 10
scale_factor_test = 100
filename = getName(scale_factor_pop, scale_factor_test)
# saveFig = 1
# prob_sick_plot_index = 4 # 4 -> 0.01
# # out of [0.001, 0.0025, 0.005, 0.0075, 0.01, 0.025, 0.05, 0.1, 0.15, 0.2, 0.25, 0.3]
# plotting(filename, prob_sick_plot_index, saveFig)
# if saveFig == 0:
# plt.show()