/
kariba_fit.py
194 lines (143 loc) · 9.42 KB
/
kariba_fit.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
import copy
from utils import is_number, val_scale, warn_p, debug_p
from sim_data_2_models import calib_data_2_models_list
from surv_data_2_ref import surv_data_2_ref as d2f
from kariba_settings import load_prevalence_mse, category_2_clusters as c2c, cluster_2_prevs as c2p, objectives_channel_codes, cluster_2_itn_traj, cluster_2_drug_cov, rdt_max
from kariba_model import KaribaModel
from kariba_utils import get_sim_group_key, get_model_params
from fitting_set import FittingSet
from fit import Fit
class KaribaFit:
def __init__(self, category, calib_data, fit_terms = None):
self.category = category
self.calib_data = calib_data
self.fit_terms = fit_terms
def fit(self):
models_list_prime = calib_data_2_models_list(self.calib_data)
best_fits = {}
all_fits = {}
#all_fits = {'fit':{'min_residual':float('inf')}, }
all_fits['min_residual'] = float('inf')
all_fits['max_residual'] = 0.0
all_fits['models'] = {}
debug_p('category ' + self.category)
for idx,cluster_id in enumerate(c2c(self.category)):
models_list = copy.deepcopy(models_list_prime)
print "Processing cluster " + cluster_id + "."
debug_p('Processing cluster ' + cluster_id + " in " + self.category + ".")
itn_traj = cluster_2_itn_traj(cluster_id)
drug_cov = cluster_2_drug_cov(cluster_id)
# prune models to the ones matching prior data
cluster_models = []
for model in models_list:
model_meta = model.get_meta()
if model_meta['group_key'] == get_sim_group_key(itn_traj, drug_cov):
#debug_p('model id before kariba conversion ' + str(model.get_model_id()))
group_key = model_meta['group_key']
sim_key = model_meta['sim_key']
model = KaribaModel(model, self.calib_data[group_key][sim_key], cluster_id, all_fits = self.fit_terms)
#model = kariba_model
#debug_p('model id after kariba conversion ' + str(model.get_model_id()))
cluster_models.append(model)
surv_data = {}
all_ref_objs_found = True
for channel_code in objectives_channel_codes:
if channel_code == 'prevalence':
prev_data = c2p(cluster_id)
if prev_data:
surv_data[channel_code] = prev_data
else:
msg = 'Prevalence objective reference data was not found!\n Skipping cluster ' + cluster_id + ' fit!'
print msg
all_ref_objs_found = False
else:
msg = "Channel objective" + channel_code + " not implemented yet!\nSetting objective reference data to None."
warn_p(msg)
surv_data[channel_code] = None
# one of the reference objective channels was not found; skipping cluster fit!
if not all_ref_objs_found:
continue
ref = d2f(surv_data)
# adjust highest possible fit to account for RDT+ model in dtk not reflecting reality at the upper end
obj_prev = ref.get_obj_by_name('prevalence')
d_points = obj_prev.get_points()
obj_prev.set_points([min(point, rdt_max) for point in d_points])
fitting_set = FittingSet(cluster_id, cluster_models, ref)
if load_prevalence_mse:
fit = Fit(fitting_set, type = 'mmse_distance_cached')
else:
fit = Fit(fitting_set)
best_fit_model = fit.best_fit_mmse_distance()
min_residual = fit.get_min_residual()
max_residual = fit.get_max_residual()
if min_residual < all_fits['min_residual']:
all_fits['min_residual'] = min_residual
if max_residual > all_fits['max_residual']:
all_fits['max_residual'] = max_residual
if best_fit_model:
temp_h, const_h, itn_level, drug_coverage_level = get_model_params(best_fit_model)
best_fit_meta = best_fit_model.get_meta()
best_fits[cluster_id] = {}
best_fits[cluster_id]['habs'] = {}
best_fits[cluster_id]['habs']['const_h'] = const_h
best_fits[cluster_id]['habs']['temp_h'] = temp_h
best_fits[cluster_id]['ITN_cov'] = itn_level
best_fits[cluster_id]['category'] = self.category
best_fits[cluster_id]['MSAT_cov'] = drug_coverage_level
best_fits[cluster_id]['sim_id'] = best_fit_meta['sim_id']
best_fits[cluster_id]['sim_key'] = best_fit_meta['sim_key']
best_fits[cluster_id]['group_key'] = best_fit_meta['group_key']
best_fits[cluster_id]['fit_value'] = best_fit_model.get_fit_val()
best_fits[cluster_id]['sim_avg_reinfection_rate'] = best_fit_model.get_sim_avg_reinfection_rate()
best_fits[cluster_id]['ref_avg_reinfection_rate'] = best_fit_model.get_ref_avg_reinfection_rate()
best_fits[cluster_id]['prevalence'] = best_fit_model.get_objective_by_name('prevalence').get_points()
# redundancy; to be refactored via FitEntry class
best_fits[cluster_id]['fit'] = {}
best_fits[cluster_id]['fit']['value'] = best_fit_model.get_fit_val()
best_fits[cluster_id]['fit']['temp_h'] = temp_h
best_fits[cluster_id]['fit']['const_h'] = const_h
best_fits[cluster_id]['fit']['ITN_cov'] = itn_level
best_fits[cluster_id]['fit']['MSAT_cov'] = drug_coverage_level
best_fits[cluster_id]['fit']['sim_id'] = best_fit_meta['sim_id']
best_fits[cluster_id]['fit']['sim_key'] = best_fit_meta['sim_key']
best_fits[cluster_id]['mse'] = {}
best_fits[cluster_id]['mse']['value'] = fit.get_min_mses()['prevalence']['value'] # get mmse for objective prevalence
best_fit_mse_model = fit.get_min_mses()['prevalence']['model']
temp_h, const_h, itn_level, drug_coverage_level = get_model_params(best_fit_mse_model)
model_meta_data = best_fit_mse_model.get_meta()
best_fits[cluster_id]['mse']['temp_h'] = temp_h
best_fits[cluster_id]['mse']['const_h'] = const_h
best_fits[cluster_id]['mse']['ITN_cov'] = itn_level
best_fits[cluster_id]['mse']['MSAT_cov'] = drug_coverage_level
best_fits[cluster_id]['mse']['sim_id'] = model_meta_data['sim_id']
best_fits[cluster_id]['mse']['sim_key'] = model_meta_data['sim_key']
best_fits[cluster_id]['cc_penalty'] = {}
best_fits[cluster_id]['cc_penalty']['value'] = fit.get_min_penalties()['prevalence']['value'] # get clinical penalty for objective prevalence; at present this is just the clinical cases penalty; if reinfection is considered the code needs to be adjusted
best_fit_cc_penalty_model = fit.get_min_penalties()['prevalence']['model']
temp_h, const_h, itn_level, drug_coverage_level = get_model_params(best_fit_cc_penalty_model)
model_meta_data = best_fit_cc_penalty_model.get_meta()
best_fits[cluster_id]['cc_penalty']['temp_h'] = temp_h
best_fits[cluster_id]['cc_penalty']['const_h'] = const_h
best_fits[cluster_id]['cc_penalty']['ITN_cov'] = itn_level
best_fits[cluster_id]['cc_penalty']['MSAT_cov'] = drug_coverage_level
best_fits[cluster_id]['cc_penalty']['sim_id'] = model_meta_data['sim_id']
best_fits[cluster_id]['cc_penalty']['sim_key'] = model_meta_data['sim_key']
rho = best_fit_model.get_rho()
p_val = best_fit_model.get_p_val()
if rho and p_val :
best_fits[cluster_id]['rho'] = rho
best_fits[cluster_id]['p_val'] = p_val
debug_p('rho' + str(rho))
debug_p('p_val' + str(p_val))
else:
msg = "something went wrong and the best fit for " + cluster_id + " could not be found."
warn_p(msg)
all_fits['models'][cluster_id] = cluster_models
#all_fits['models'][cluster_id] = fit.get_fitting_set_models()
print str(idx+1) + " clusters have been processed."
debug_p( str(idx+1) + " clusters have been processed in category " + self.category)
'''
if idx > 0:
break
'''
return best_fits, all_fits