gluonts.dataset.artificial.recipe module#
- class gluonts.dataset.artificial.recipe.ARp(phi: Union[Any, Callable], sigma: Union[Any, Callable], xhist: Optional[Union[Any, Callable]] = None, c: Union[Any, Callable] = 0.0, noise: Optional[Union[Any, Callable]] = None)[source]#
- class gluonts.dataset.artificial.recipe.BinaryHolidays(dates: List[pandas._libs.tslibs.timestamps.Timestamp], holidays: List[Any])[source]#
- class gluonts.dataset.artificial.recipe.BinaryMarkovChain(one_to_zero: Union[Any, Callable], zero_to_one: Union[Any, Callable])[source]#
- class gluonts.dataset.artificial.recipe.Choose(options: Union[Any, Callable], selector: Union[Any, Callable])[source]#
- class gluonts.dataset.artificial.recipe.Concatenate(inputs: List[Union[Any, Callable]], axis: int = 0)[source]#
- class gluonts.dataset.artificial.recipe.Convolve(input: Union[Any, Callable], filter: Union[Any, Callable])[source]#
- class gluonts.dataset.artificial.recipe.EvalRecipe(recipe: Union[Callable, List[Callable], List[Tuple[str, Callable]], Dict[str, Callable]], op: Optional[Union[Any, Callable]] = None)[source]#
- class gluonts.dataset.artificial.recipe.Lag(input: Union[Any, Callable], lag: Union[Any, Callable] = 0, pad_const: int = 0)[source]#
- class gluonts.dataset.artificial.recipe.NanWhere(source: Union[Any, Callable], nan_indicator: Union[Any, Callable])[source]#
- class gluonts.dataset.artificial.recipe.NumpyFunc(func: str, func_args: Tuple[Any, ...], func_kwargs: Dict[str, Any])[source]#
- class gluonts.dataset.artificial.recipe.RandomCat(cardinalities: List[int], prob_fun: Callable = gluonts.dataset.artificial.recipe.RandomSymmetricDirichlet(alpha=1.0, shape=(0,)))[source]#
Bases:
object
- class gluonts.dataset.artificial.recipe.RandomChangepoints(max_num_changepoints: Union[Any, Callable])[source]#
- class gluonts.dataset.artificial.recipe.RandomGaussian(stddev: Union[Any, Callable] = 1.0, shape: Sequence[int] = (0,))[source]#
- class gluonts.dataset.artificial.recipe.RandomInteger(low: Union[Any, Callable], high: Union[Any, Callable], shape: Optional[Sequence[int]] = (0,))[source]#
- class gluonts.dataset.artificial.recipe.RandomSymmetricDirichlet(alpha: Union[Any, Callable] = 1.0, shape: Sequence[int] = (0,))[source]#
- class gluonts.dataset.artificial.recipe.RandomUniform(low: Union[Any, Callable] = 0.0, high: Union[Any, Callable] = 1.0, shape=(0,))[source]#
- class gluonts.dataset.artificial.recipe.SmoothSeasonality(period: Union[Any, Callable], phase: Union[Any, Callable])[source]#
- gluonts.dataset.artificial.recipe.evaluate(recipe: Union[Callable, List[Callable], List[Tuple[str, Callable]], Dict[str, Callable]], length: Union[Any, Callable], *args, global_state: Optional[dict] = None, **kwargs) Any[source]#
- gluonts.dataset.artificial.recipe.generate(length: int, recipe: Union[Callable, List[Callable], List[Tuple[str, Callable]], Dict[str, Callable]], start: pandas._libs.tslibs.period.Period, global_state: Optional[dict] = None, seed: int = 0, item_id_prefix: str = '') Iterator[Dict[str, Any]][source]#
- gluonts.dataset.artificial.recipe.lift(input: Union[int, Callable])[source]#
Use this decorator to lift a function.
@lift def f(x, y, length=None)
or if your function returns more results
@lift(2) def f(x, y, length=None)
You can then use your function as part of a recipe. The function is called with all all arguments being already resolved.
Note that you cannot serialize recipes that use the lift decorated functions.
- gluonts.dataset.artificial.recipe.make_func(length: int, recipe: Union[Callable, List[Callable], List[Tuple[str, Callable]], Dict[str, Callable]], global_state=None) Callable[[int, Dict[str, Any]], Dict[str, Any]][source]#
- gluonts.dataset.artificial.recipe.normalized_ar1(tau, x0=None, norm='minmax', sigma=1.0)[source]#
Returns an ar1 process with an auto correlation time of tau.
- norm can be:
None -> no normalization
‘minmax’ -> min_max_scaled
‘standard’ -> 0 mean, unit variance