MODULE REFERENCE

datasets

anomx.integrations.darts.datasets

Darts-native datasets and optional, platform-agnostic channel loading.

Public exports

Import these names directly from anomx.integrations.darts.datasets.

AnomxDataset

An Anomx time series that is directly usable by every Darts model.

DartsDataset

An Anomx time series that is directly usable by every Darts model.

classView source ↗

Dataset

Dataset

Bases: TimeSeries

An Anomx time series that is directly usable by every Darts model.

All Darts factories, splitting, transforms, slicing, and covariate APIs are available. from_channels additionally resolves Anomx object references or external references (for example DOOCS properties) through a host loader or an explicit anomx.integrations.platform.PlatformClient.

Members

from_channelsclassmethod

# Dataset.from_channels Source ↗

from_channels(channels: Sequence[str], *, client: ChannelClient | PlatformClient | None = None, loader: Callable[..., Any] | None = None, start: str | None = None, end: str | None = None, frequency: str | None = None, fill_missing_dates: bool = False) -> Dataset

Resolve references into aligned, numeric, regularly sampled data.

Loaders receive channels, start, end, and frequency as keyword arguments and return a TimeSeries or wide pandas DataFrame. Frames use a DatetimeIndex or a timestamp column. Duplicate times, missing values and unaligned channels fail explicitly. This class does not fill or average values; the loader defines any preprocessing.

Signature parameters
Parameter	Type	Default
`channels`	`Sequence[str]`	`required`
`client`keyword only	`ChannelClient \| PlatformClient \| None`	`None`
`loader`keyword only	`Callable[..., Any] \| None`	`None`
`start`keyword only	`str \| None`	`None`
`end`keyword only	`str \| None`	`None`
`frequency`keyword only	`str \| None`	`None`
`fill_missing_dates`keyword only	`bool`	`False`

Returns Dataset