def fit_ensemble(self,
task_id=None,
hdl_id=None,
trials_fetcher="GetBestK",
trials_fetcher_params=frozendict(k=10),
ensemble_type="stack",
ensemble_params=frozendict(),
return_Xy_test=False):
if task_id is None:
assert hasattr(
self.resource_manager,
"task_id") and self.resource_manager.task_id is not None
task_id = self.resource_manager.task_id
# if hdl_id is None:
# assert hasattr(self.resource_manager, "hdl_id") and self.resource_manager.hdl_id is not None
# hdl_id = self.resource_manager.hdl_id
trials_fetcher_name = trials_fetcher
from autoflow.ensemble import trials_fetcher
assert hasattr(trials_fetcher, trials_fetcher_name)
trials_fetcher_cls = getattr(trials_fetcher, trials_fetcher_name)
trials_fetcher: TrialsFetcher = trials_fetcher_cls(
resource_manager=self.resource_manager,
task_id=task_id,
hdl_id=hdl_id,
**trials_fetcher_params)
trial_ids = trials_fetcher.fetch()
estimator_list, y_true_indexes_list, y_preds_list = TrainedDataFetcher(
task_id, hdl_id, trial_ids, self.resource_manager).fetch()
ml_task, Xy_train, Xy_test = self.resource_manager.get_ensemble_needed_info(
task_id, hdl_id)
y_true = Xy_train[1]
ensemble_estimator_package_name = f"autoflow.ensemble.{ensemble_type}.{ml_task.role}"
ensemble_estimator_package = import_module(
ensemble_estimator_package_name)
ensemble_estimator_class_name = get_class_name_of_module(
ensemble_estimator_package_name)
ensemble_estimator_class = getattr(ensemble_estimator_package,
ensemble_estimator_class_name)
ensemble_estimator: EnsembleEstimator = ensemble_estimator_class(
**ensemble_params)
ensemble_estimator.fit_trained_data(estimator_list,
y_true_indexes_list, y_preds_list,
y_true)
self.ensemble_estimator = ensemble_estimator
if return_Xy_test:
return self.ensemble_estimator, Xy_test
else:
return self.ensemble_estimator
def get_forbid_hit_in_models_by_rely(self,
models,
rely_model="boost_model"):
forbid_in_value = []
hit = []
for model in models:
module_path = f"autoflow.pipeline.components.{self.ml_task.mainTask}.{model}"
_class = get_class_name_of_module(module_path)
M = import_module(module_path)
cls = getattr(M, _class)
is_hit = getattr(cls, rely_model, False)
if not is_hit:
forbid_in_value.append(model)
else:
hit.append(model)
return forbid_in_value, hit
def fit_ensemble(self,
task_id=None,
hdl_id=None,
trials_fetcher_cls="GetBestK",
trials_fetcher_params=frozendict(k=10),
ensemble_type="stack",
ensemble_params=frozendict(),
fit_ensemble_alone=True):
# fixme: ensemble_params可能会面临一个问题，就是传入无法序列化的内容
trials_fetcher_params = dict(trials_fetcher_params)
ensemble_params = dict(ensemble_params)
kwargs = get_valid_params_in_kwargs(self.fit_ensemble, locals())
if task_id is None:
assert hasattr(
self.resource_manager,
"task_id") and self.resource_manager.task_id is not None
task_id = self.resource_manager.task_id
self.task_id = task_id
self.resource_manager.task_id = task_id
if hdl_id is not None:
self.hdl_id = hdl_id
self.resource_manager.hdl_id = hdl_id
if fit_ensemble_alone:
setup_logger(self.log_path, self.log_config)
if fit_ensemble_alone:
experiment_config = {"fit_ensemble_params": kwargs}
self.resource_manager.insert_experiment_record(
ExperimentType.ENSEMBLE, experiment_config, {})
self.experiment_id = self.resource_manager.experiment_id
from autoflow.ensemble import trials_fetcher
assert hasattr(trials_fetcher, trials_fetcher_cls)
trials_fetcher_cls = getattr(trials_fetcher, trials_fetcher_cls)
trials_fetcher_inst: TrialsFetcher = trials_fetcher_cls(
resource_manager=self.resource_manager,
task_id=task_id,
hdl_id=hdl_id,
**trials_fetcher_params)
trial_ids = trials_fetcher_inst.fetch()
estimator_list, y_true_indexes_list, y_preds_list = TrainedDataFetcher(
task_id, hdl_id, trial_ids, self.resource_manager).fetch()
# todo: 在这里，只取了验证集的数据，没有取测试集的数据。待拓展
ml_task, y_true = self.resource_manager.get_ensemble_needed_info(
task_id)
if len(estimator_list) == 0:
raise ValueError("Length of estimator_list must >=1. ")
elif len(estimator_list) == 1:
self.logger.info(
"Length of estimator_list == 1, don't do ensemble.")
if ml_task.mainTask == "classification":
ensemble_estimator = VoteClassifier(estimator_list[0])
else:
ensemble_estimator = MeanRegressor(estimator_list[0])
else:
ensemble_estimator_package_name = f"autoflow.ensemble.{ensemble_type}.{ml_task.role}"
ensemble_estimator_package = import_module(
ensemble_estimator_package_name)
ensemble_estimator_class_name = get_class_name_of_module(
ensemble_estimator_package_name)
ensemble_estimator_class = getattr(ensemble_estimator_package,
ensemble_estimator_class_name)
# ensemble_estimator : EnsembleEstimator
ensemble_estimator = ensemble_estimator_class(**ensemble_params)
ensemble_estimator.fit_trained_data(estimator_list,
y_true_indexes_list,
y_preds_list, y_true)
self.ensemble_estimator = ensemble_estimator
if fit_ensemble_alone:
self.estimator = self.ensemble_estimator
self.resource_manager.finish_experiment(self.log_path, self)
return self.ensemble_estimator