-
Notifications
You must be signed in to change notification settings - Fork 0
/
simulation_functions.py
168 lines (114 loc) · 5.69 KB
/
simulation_functions.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
from math import ceil
import cupy as cp
import matplotlib.pyplot as plt
import numpy as np
from phaseportrait.sliders import Slider
from .evolution import *
from .parameters_control import *
from .configuration import smooth_deaths_list
epi_poblation_to_index = {
's' : 0,
'e' : 1,
'i' : 2,
'd' : 3,
}
def evolve_gpu(params, fixed_params, state, deaths_list, log_diff, configuration):
total_population = configuration["total_population"]
max_days = configuration["max_days"]
time = cp.zeros(params.shape[1])
deaths_ref = cp.zeros(params.shape[1])
_time = cp.zeros(params.shape[1], dtype=cp.int32)
_time[:] = time[:]
# max_deaths_ref = max(deaths_list)
__time_ref = 0
# while (_time<max_days).any():
while __time_ref<max_days:
evolve(params, fixed_params, state)
deaths = state[3]*total_population
# deaths_ref = 0 + deaths_list[_time * (_time>=0)] #TODO: mirar si esto hace falta * (_time>=0)
deaths_ref = 0 + deaths_list[__time_ref]
diff = cp.abs(deaths-deaths_ref)
log_diff += cp.log(diff+1)
_time+=1
__time_ref+=1
def evolve_gpu_no_diff(params, fixed_params, state, p_active, max_days=1):
time = 0
while time<max_days:
evolve(params, fixed_params, state, time, p_active)
time+=1
###############################################
## Seleccionar los 5% mejores
def get_best_parameters(params, log_diff, save_percentage):
"Retuns the best `save_percentage`% `params` of the simulations given their `log_diff` with real data."
log_diff_index_sorted = cp.argsort(log_diff)
# Para comprobar que indices tomar en el sort
# print(log_diff[log_diff_index_sorted[0]], log_diff[log_diff_index_sorted[-1]])
save_count = ceil(log_diff.size*save_percentage*0.01)
# save_count = 6
saved_params = cp.zeros((len(param_to_index),save_count), dtype=cp.float64)
saved_log_diff = cp.zeros(save_count, dtype=cp.float64)
for i in range(save_count):
saved_params[:,i] = params[:,log_diff_index_sorted[i]]
saved_log_diff[i] = log_diff[log_diff_index_sorted[i]]
return saved_params, saved_log_diff
class Simulation:
_name_='Simulation_PhasePortrait'
def add_slider(self, param_name, *, valinit=None, valstep=0.1, valinterval=10):
self.sliders.update({param_name: Slider(self, param_name, valinit=valinit, valstep=valstep, valinterval=valinterval)})
self.fig.subplots_adjust(bottom=0.25)
self.sliders[param_name].slider.on_changed(self.sliders[param_name])
def __init__(self, deaths_list, p_active, fixed_params, max_days=1, total_population=1) -> None:
self.fig, self.ax = plt.subplots()
self.fixed_params = fixed_params
self.max_days = max_days
self.total_population = total_population
self.parameters = np.zeros(len(param_to_index), dtype=np.float64)
self.p_active = p_active
self.deaths_list = deaths_list
self.sliders = {}
self.ax2 = self.ax.twinx()
def plot(self):
for p, s in self.sliders.items():
try:
self.parameters[param_to_index[p]] = s.value
except KeyError:
self.fixed_params[fixed_params_to_index[p]] = s.value
self.state = np.zeros(7, dtype=np.float64)
self.state[1] = 1- self.parameters[param_to_index['initial_i']]
self.state[3] = self.parameters[param_to_index['initial_i']]
time_list = range(len(self.deaths_list)+1)
self.ax.plot(time_list[:self.max_days], smooth_deaths_list(self.deaths_list[:self.max_days]).get() , color='purple', label='smooth data')
self.ax.plot(time_list[:self.max_days], self.deaths_list[:self.max_days], label='real data', color='red')
deaths_list = np.zeros(self.max_days)
log_diff_hist = np.zeros(self.max_days)
time = int(self.parameters[param_to_index['offset']])
__time_ref = 0
log_diff = 0
# print(time)
while time<self.max_days:
# while __time_ref<max_days:
evolve(self.parameters, self.fixed_params, self.state, self.p_active[__time_ref])
deaths = self.state[5]*self.total_population
deaths_ref = 0 + self.deaths_list[time * (time>=0)] #TODO: mirar si esto hace falta * (time>=0)
# deaths_ref = 0 + deaths_list[__time_ref]
if __time_ref < self.max_days:
deaths_list[__time_ref] = self.state[5]*self.total_population
# diff = (deaths/(deaths_ref + 1*(deaths_ref<1)))
diff = cp.abs(deaths - deaths_ref)
# diff = cp.abs(cp.log(deaths+1)-cp.log(deaths_ref+1))
log_diff_hist[__time_ref] = cp.abs(cp.log(diff + 1))
# log_diff_hist[__time_ref] = (diff) * (time<self.max_days)
log_diff += log_diff_hist[__time_ref]
time+=1
__time_ref+=1
self.ax.plot(time_list[:self.max_days], deaths_list, '-', color='black', label=f'{log_diff}')
self.ax.set_title('Muertes diarias (España)')
self.ax2.clear()
self.ax2.plot(time_list[:self.max_days], log_diff_hist, color='green')
self.ax.set_yscale('log')
self.ax.legend()
# self.ax.set_ylim(0, max(self.deaths_list[:self.max_days]))
# self.ax2 = self.ax.twinx()
# self.ax2.plot(time_list[:self.max_days], self.p_active[:self.max_days], color='green', label='p_active')
# self.ax2.tick_params(axis ='y', labelcolor = 'green')
# self.ax2.legend()