def run_ensemble(self):
"""
Run an ensemble of models for each suppression policy nad generate the
output report / results dataset.
"""
for suppression_policy_name, suppression_policy in self.suppression_policies.items():
_logger.info(f'Running simulation ensemble for {self.state_name} {self.fips} {suppression_policy_name}')
if self.run_mode is RunMode.CAN_INFERENCE_DERIVED:
model_ensemble = [self._load_model_for_fips(scenario=suppression_policy)]
else:
raise ValueError(f'Run mode {self.run_mode.value} not supported.')
if self.agg_level is AggregationLevel.COUNTY:
self.all_outputs['county_metadata'] = self.county_metadata
self.all_outputs['county_metadata']['age_distribution'] = list(self.all_outputs['county_metadata']['age_distribution'])
self.all_outputs['county_metadata']['age_bins'] = list(self.all_outputs['county_metadata']['age_distribution'])
self.all_outputs[f'{suppression_policy_name}'] = self._generate_output_for_suppression_policy(model_ensemble)
if self.generate_report and self.output_file_report:
report = CountyReport(self.fips,
model_ensemble=model_ensemble,
county_outputs=self.all_outputs,
filename=self.output_file_report,
summary=self.summary)
report.generate_and_save()
with open(self.output_file_data, 'w') as f:
json.dump(self.all_outputs, f)
def run_ensemble(self):
"""
Run an ensemble of models for each suppression policy nad generate the
output report / results dataset.
"""
for suppression_policy_name, suppression_policy in self.suppression_policies.items(
):
logging.info(
f'Running simulation ensemble for {self.state_name} {self.fips} {suppression_policy_name}'
)
if suppression_policy_name == 'suppression_policy__inferred':
artifact_path = get_run_artifact_path(
self.fips, RunArtifact.MLE_FIT_MODEL)
if os.path.exists(artifact_path):
with open(artifact_path, 'rb') as f:
model_ensemble = [pickle.load(f)]
else:
logging.warning(
f'No MLE model found for {self.state_name}: {self.fips}. Skipping.'
)
else:
parameter_sampler = ParameterEnsembleGenerator(
fips=self.fips,
N_samples=self.n_samples,
t_list=self.t_list,
suppression_policy=suppression_policy)
parameter_ensemble = parameter_sampler.sample_seir_parameters(
override_params=self.override_params)
model_ensemble = list(
map(self._run_single_simulation, parameter_ensemble))
if self.agg_level is AggregationLevel.COUNTY:
self.all_outputs['county_metadata'] = self.county_metadata
self.all_outputs['county_metadata']['age_distribution'] = list(
self.all_outputs['county_metadata']['age_distribution'])
self.all_outputs['county_metadata']['age_bins'] = list(
self.all_outputs['county_metadata']['age_distribution'])
self.all_outputs[
f'{suppression_policy_name}'] = self._generate_output_for_suppression_policy(
model_ensemble)
if self.generate_report and self.output_file_report:
report = CountyReport(self.fips,
model_ensemble=model_ensemble,
county_outputs=self.all_outputs,
filename=self.output_file_report,
summary=self.summary)
report.generate_and_save()
with open(self.output_file_data, 'w') as f:
json.dump(self.all_outputs, f)
def run_ensemble(self):
"""
Run an ensemble of models for each suppression policy nad generate the
output report / results dataset.
"""
for suppression_policy in self.suppression_policy:
logging.info(f'Generating For Policy {suppression_policy}')
parameter_ensemble = ParameterEnsembleGenerator(
fips=self.fips,
N_samples=self.n_samples,
t_list=self.t_list,
suppression_policy=generate_empirical_distancing_policy(
t_list=self.t_list,
fips=self.fips,
future_suppression=suppression_policy)
).sample_seir_parameters()
model_ensemble = list(
map(self._run_single_simulation, parameter_ensemble))
logging.info(
f'Generating Report for suppression policy {suppression_policy}'
)
self.all_outputs[f'suppression_policy__{suppression_policy}'] = \
self._generate_output_for_suppression_policy(model_ensemble, suppression_policy)
if self.generate_report:
report = CountyReport(self.fips,
model_ensemble=model_ensemble,
county_outputs=self.all_outputs,
filename=self.output_file_report,
summary=self.summary)
report.generate_and_save()
with open(self.output_file_data, 'w') as f:
json.dump(self.all_outputs, f)