modified methods to get conditional probability data
This commit is contained in:
lrdossan
parent
1ddbf64de3
commit
eb0f5cc881
2 changed files with 34 additions and 38 deletions
|
|
@ -171,26 +171,15 @@ def calculate_report_data(form: VirusFormData, model: models.ExposureModel, exec
|
|||
expected_new_cases = np.array(model.expected_new_cases()).mean()
|
||||
exposed_presence_intervals = [list(interval) for interval in model.exposed.presence_interval().boundaries()]
|
||||
|
||||
if (model.data_registry.virological_data['virus_distributions'][form.virus_type]['viral_load_in_sputum'] == ViralLoads.COVID_OVERALL.value # type: ignore
|
||||
and form.conditional_probability_plot): # Only generate this data if covid_overall_vl_data is selected.
|
||||
|
||||
viral_load_in_sputum: models._VectorisedFloat = model.concentration_model.infected.virus.viral_load_in_sputum
|
||||
viral_loads, pi_means, lower_percentiles, upper_percentiles = manufacture_conditional_probability_data(model, prob)
|
||||
|
||||
uncertainties_plot_src = img2base64(_figure2bytes(uncertainties_plot(prob, viral_load_in_sputum, viral_loads,
|
||||
pi_means, lower_percentiles, upper_percentiles)))
|
||||
conditional_probability_data = {key: value for key, value in
|
||||
zip(('viral_loads', 'pi_means', 'lower_percentiles', 'upper_percentiles'),
|
||||
(viral_loads, pi_means, lower_percentiles, upper_percentiles))}
|
||||
vl_dist = list(np.log10(viral_load_in_sputum))
|
||||
|
||||
else:
|
||||
uncertainties_plot_src = None
|
||||
conditional_probability_data = None
|
||||
vl = model.concentration_model.virus.viral_load_in_sputum
|
||||
if isinstance(vl, np.ndarray): vl_dist = list(np.log10(model.concentration_model.virus.viral_load_in_sputum))
|
||||
else: vl_dist = np.log10(model.concentration_model.virus.viral_load_in_sputum)
|
||||
|
||||
conditional_probability_data = None
|
||||
uncertainties_plot_src = None
|
||||
if (form.conditional_probability_viral_loads and
|
||||
model.data_registry.virological_data['virus_distributions'][form.virus_type]['viral_load_in_sputum'] == ViralLoads.COVID_OVERALL.value): # type: ignore
|
||||
# Generate all the required data for the conditional probability plot
|
||||
conditional_prob_data = manufacture_conditional_probability_data(model, prob)
|
||||
# Generate the matplotlib image based on the received data
|
||||
uncertainties_plot_src = img2base64(_figure2bytes(uncertainties_plot(prob, conditional_prob_data)))
|
||||
|
||||
return {
|
||||
"model_repr": repr(model),
|
||||
"times": list(times),
|
||||
|
|
@ -208,10 +197,9 @@ def calculate_report_data(form: VirusFormData, model: models.ExposureModel, exec
|
|||
"prob_hist_bins": list(prob_dist_bins),
|
||||
"prob_probabilistic_exposure": prob_probabilistic_exposure,
|
||||
"expected_new_cases": expected_new_cases,
|
||||
"uncertainties_plot_src": uncertainties_plot_src,
|
||||
"CO2_concentrations": CO2_concentrations,
|
||||
"vl_dist": vl_dist,
|
||||
"conditional_probability_data": conditional_probability_data,
|
||||
"uncertainties_plot_src": uncertainties_plot_src,
|
||||
}
|
||||
|
||||
|
||||
|
|
@ -235,7 +223,6 @@ def generate_permalink(base_url, get_root_url, get_root_calculator_url, form: V
|
|||
|
||||
|
||||
def conditional_prob_inf_given_vl_dist(
|
||||
data_registry: DataRegistry,
|
||||
infection_probability: models._VectorisedFloat,
|
||||
viral_loads: np.ndarray,
|
||||
specific_vl: float,
|
||||
|
|
@ -247,39 +234,48 @@ def conditional_prob_inf_given_vl_dist(
|
|||
upper_percentiles = []
|
||||
|
||||
for vl_log in viral_loads:
|
||||
# Viral load corresponding to a certain viral load value in the distribution
|
||||
specific_prob = infection_probability[np.where((vl_log-step/2-specific_vl)*(vl_log+step/2-specific_vl)<0)[0]] #type: ignore
|
||||
|
||||
pi_means.append(specific_prob.mean())
|
||||
lower_percentiles.append(np.quantile(specific_prob, 0.05))
|
||||
upper_percentiles.append(np.quantile(specific_prob, 0.95))
|
||||
|
||||
return pi_means, lower_percentiles, upper_percentiles
|
||||
return np.array(pi_means), np.array(lower_percentiles), np.array(upper_percentiles)
|
||||
|
||||
|
||||
def manufacture_conditional_probability_data(
|
||||
exposure_model: models.ExposureModel,
|
||||
infection_probability: models._VectorisedFloat
|
||||
infection_probability: models._VectorisedFloat,
|
||||
):
|
||||
data_registry: DataRegistry = exposure_model.data_registry
|
||||
|
||||
min_vl = 2
|
||||
max_vl = 10
|
||||
step = (max_vl - min_vl)/100
|
||||
viral_loads = np.arange(min_vl, max_vl, step)
|
||||
specific_vl = np.log10(exposure_model.concentration_model.virus.viral_load_in_sputum)
|
||||
pi_means, lower_percentiles, upper_percentiles = conditional_prob_inf_given_vl_dist(data_registry, infection_probability, viral_loads,
|
||||
pi_means, lower_percentiles, upper_percentiles = conditional_prob_inf_given_vl_dist(infection_probability, viral_loads,
|
||||
specific_vl, step)
|
||||
|
||||
return list(viral_loads), list(pi_means), list(lower_percentiles), list(upper_percentiles)
|
||||
log10_vl_in_sputum : models._VectorisedFloat = np.log10(exposure_model.concentration_model.infected.virus.viral_load_in_sputum)
|
||||
|
||||
return {
|
||||
'viral_loads': viral_loads,
|
||||
'pi_means': pi_means,
|
||||
'lower_percentiles': lower_percentiles,
|
||||
'upper_percentiles': upper_percentiles,
|
||||
'log10_vl_in_sputum': log10_vl_in_sputum,
|
||||
}
|
||||
|
||||
|
||||
def uncertainties_plot(infection_probability: models._VectorisedFloat,
|
||||
viral_load_in_sputum: models._VectorisedFloat,
|
||||
viral_loads: models._VectorisedFloat,
|
||||
pi_means: models._VectorisedFloat,
|
||||
lower_percentiles: models._VectorisedFloat,
|
||||
upper_percentiles: models._VectorisedFloat):
|
||||
conditional_prob_data: dict):
|
||||
|
||||
fig, axes = plt.subplots(2, 3,
|
||||
viral_loads: models._VectorisedFloat = conditional_prob_data['viral_loads']
|
||||
pi_means: models._VectorisedFloat = conditional_prob_data['pi_means']
|
||||
lower_percentiles: models._VectorisedFloat = conditional_prob_data['lower_percentiles']
|
||||
upper_percentiles: models._VectorisedFloat = conditional_prob_data['upper_percentiles']
|
||||
log10_vl_in_sputum: models._VectorisedFloat = conditional_prob_data['log10_vl_in_sputum']
|
||||
|
||||
fig, axes = plt.subplots(2, 3,
|
||||
gridspec_kw={'width_ratios': [5, 0.5] + [1],
|
||||
'height_ratios': [3, 1], 'wspace': 0},
|
||||
sharey='row',
|
||||
|
|
@ -306,7 +302,7 @@ def uncertainties_plot(infection_probability: models._VectorisedFloat,
|
|||
|
||||
axs[0, 2].text(highest_bar * 0.5, 50,
|
||||
"$P(I)=$\n" + rf"$\bf{np.round(np.mean(infection_probability), 1)}$%", ha='center', va='center')
|
||||
axs[1, 0].hist(np.log10(viral_load_in_sputum),
|
||||
axs[1, 0].hist(log10_vl_in_sputum,
|
||||
bins=150, range=(2, 10), color='grey')
|
||||
axs[1, 0].set_facecolor("lightgrey")
|
||||
axs[1, 0].set_yticks([])
|
||||
|
|
|
|||
|
|
@ -72,7 +72,7 @@ def test_conditional_prob_inf_given_vl_dist(data_registry, baseline_exposure_mod
|
|||
specific_vl = np.log10(mc_model.concentration_model.infected.virus.viral_load_in_sputum)
|
||||
step = 8/100
|
||||
actual_pi_means, actual_lower_percentiles, actual_upper_percentiles = (
|
||||
report_generator.conditional_prob_inf_given_vl_dist(data_registry, infection_probability, viral_loads, specific_vl, step)
|
||||
report_generator.conditional_prob_inf_given_vl_dist(infection_probability, viral_loads, specific_vl, step)
|
||||
)
|
||||
|
||||
assert np.allclose(actual_pi_means, expected_pi_means, atol=0.002)
|
||||
|
|
|
|||
Loading…
Reference in a new issue