forecast_command

Functions

compute_horizon(horizon, forecast_params, ...)	Compute new pipeline horizon if start_timestamp presented in forecast_params.
filter_forecast(forecast_ts, forecast_params)	Filter out forecasts before start_timestamp if start_timestamp presented in forecast_params..
forecast([config_path, target_path, freq, ...])	Command to make forecast with etna without coding.
update_horizon(pipeline_configs, ...)	Update the horizon parameter in the pipeline config if start_timestamp is set.

compute_horizon(horizon: int, forecast_params: Dict[str, Any], tsdataset: etna.datasets.tsdataset.TSDataset) → int

Compute new pipeline horizon if start_timestamp presented in forecast_params.

Parameters

• horizon (int) –

• forecast_params (Dict[str, Any]) –

• tsdataset (etna.datasets.tsdataset.TSDataset) –

Return type

int

filter_forecast(forecast_ts: etna.datasets.tsdataset.TSDataset, forecast_params: Dict[str, Any]) → etna.datasets.tsdataset.TSDataset

Filter out forecasts before start_timestamp if start_timestamp presented in forecast_params..

Parameters

• forecast_ts (etna.datasets.tsdataset.TSDataset) –

• forecast_params (Dict[str, Any]) –

Return type

etna.datasets.tsdataset.TSDataset

forecast(config_path: pathlib.Path = <typer.models.ArgumentInfo object>, target_path: pathlib.Path = <typer.models.ArgumentInfo object>, freq: str = <typer.models.ArgumentInfo object>, output_path: pathlib.Path = <typer.models.ArgumentInfo object>, exog_path: typing.Optional[pathlib.Path] = <typer.models.ArgumentInfo object>, forecast_config_path: typing.Optional[pathlib.Path] = <typer.models.ArgumentInfo object>, raw_output: bool = <typer.models.ArgumentInfo object>, known_future: typing.Optional[typing.List[str]] = <typer.models.ArgumentInfo object>)

Command to make forecast with etna without coding.

Expected format of csv with target timeseries:

 ============= =========== ==========

timestamp segment target

2020-01-01	segment_1	1
2020-01-02	segment_1	2
2020-01-03	segment_1	3
2020-01-04	segment_1	4
…
2020-01-10	segment_2	10
2020-01-11	segment_2	20

Expected format of csv with exogenous timeseries:

 ============= =========== =============== ===============

timestamp segment regressor_1 regressor_2

2020-01-01	segment_1	11	12
2020-01-02	segment_1	22	13
2020-01-03	segment_1	31	14
2020-01-04	segment_1	42	15
…
2020-02-10	segment_2	101	61
2020-02-11	segment_2	205	54

Parameters

• config_path (pathlib.Path) –

• target_path (pathlib.Path) –

• freq (str) –

• output_path (pathlib.Path) –

• exog_path (Optional[pathlib.Path]) –

• forecast_config_path (Optional[pathlib.Path]) –

• raw_output (bool) –

• known_future (Optional[List[str]]) –

update_horizon(pipeline_configs: Dict[str, Any], forecast_params: Dict[str, Any], tsdataset: etna.datasets.tsdataset.TSDataset)

Update the horizon parameter in the pipeline config if start_timestamp is set.

Parameters

• pipeline_configs (Dict[str, Any]) –

• forecast_params (Dict[str, Any]) –

• tsdataset (etna.datasets.tsdataset.TSDataset) –