def train_for_region(self, data_source, region_type, region_name, train_start_date, train_end_date,
search_space, search_parameters, train_loss_function, input_filepath):
observations = DataFetcherModule.get_observations_for_region(region_type, region_name,
data_source=data_source, filepath=input_filepath)
region_metadata = DataFetcherModule.get_regional_metadata(region_type, region_name, data_source=data_source)
return self.train(region_metadata, observations, train_start_date, train_end_date,
search_space, search_parameters, train_loss_function)
def predict_for_region(self, region_type, region_name, run_day, forecast_start_date,
forecast_end_date):
observations = DataFetcherModule.get_observations_for_region(region_type, region_name)
region_metadata = DataFetcherModule.get_regional_metadata(region_type, region_name)
return self.predict(region_type, region_name, region_metadata, observations, run_day,
forecast_start_date,
forecast_end_date)
def evaluate_for_region(self, region_type, region_name, run_day,
test_start_date, test_end_date, loss_functions):
observations = DataFetcherModule.get_observations_for_region(
region_type, region_name)
region_metadata = DataFetcherModule.get_regional_metadata(
region_type, region_name)
return self.evaluate(region_metadata, observations, run_day,
test_start_date, test_end_date, loss_functions)
def train_for_region(self, region_type, region_name, train_start_date,
train_end_date, search_space, search_parameters,
train_loss_function):
observations = DataFetcherModule.get_observations_for_region(
region_type, region_name)
region_metadata = DataFetcherModule.get_regional_metadata(
region_type, region_name)
return self.train(region_metadata, observations, train_start_date,
train_end_date, search_space, search_parameters,
train_loss_function)
def evaluate_for_region(self, data_source, region_type, region_name,
run_day, test_start_date, test_end_date,
loss_functions, input_filepath):
observations = DataFetcherModule.get_observations_for_region(
region_type,
region_name,
data_source=data_source,
filepath=input_filepath)
region_metadata = DataFetcherModule.get_regional_metadata(
region_type, region_name, data_source=data_source)
return self.evaluate(region_metadata, observations, run_day,
test_start_date, test_end_date, loss_functions)
def predict_for_region(self, data_source, region_type, region_name,
run_day, forecast_start_date, forecast_end_date,
input_filepath):
observations = DataFetcherModule.get_observations_for_region(
region_type,
region_name,
data_source=data_source,
filepath=input_filepath)
region_metadata = DataFetcherModule.get_regional_metadata(
region_type, region_name, data_source=data_source)
return self.predict(region_type, region_name, region_metadata,
observations, run_day, forecast_start_date,
forecast_end_date)
def predict_for_region(self, data_source: DataSource, region_type: str, region_name: List[str], run_day: str,
start_date: str, input_type: InputType, time_intervals: List[ForecastTimeInterval],
input_filepath: str):
"""
method downloads data using data fetcher module and then run predict on that dataset.
@param region_type: region_type supported by data_fetcher module
@param region_name: region_name supported by data_fetcher module
@param run_day: date of initialization
@param start_date: start_date
@param input_type: input_type can be npi_list/param_override
@param time_intervals: list of time_intervals with parameters
@param data_source: data source
@param input_filepath: input data file path
@return: pd.DataFrame: predictions
"""
observations = DataFetcherModule.get_observations_for_region(region_type, region_name, data_source=data_source,
filepath=input_filepath)
region_metadata = DataFetcherModule.get_regional_metadata(region_type, region_name)
return self.predict(region_type, region_name, region_metadata, observations, run_day,
start_date, input_type, time_intervals)