Module transform

Dataset Transformations

This module provides a collection of crate::Dataset composition wrappers; providing composition, subset selection, sampling, random shuffling, and windowing.

• ComposedDataset - composes a list of datasets.
• PartialDataset - selects a contiguous index range subset of a dataset.
• ShuffledDataset - a randomly shuffled / mutably shuffle-able dataset; a thin wrapper around SelectionDataset.
• SamplerDataset - samples a dataset; support for with/without replacement, and under/oversampling.
• SelectionDataset - selects a subset of a dataset via indices; support for shuffling.
• WindowsDataset - creates a sliding window over a dataset.

Structs

ComposedDataset

Compose multiple datasets together to create a bigger one.

MapperDataset

Dataset mapping each element in an inner dataset to another element type lazily.

PartialDataset

Only use a fraction of an existing dataset lazily.

SamplerDataset

Sample items from a dataset.

SamplerDatasetOptions

Options to configure a SamplerDataset.

SelectionDataset

A dataset that selects a subset of indices from an existing dataset.

ShuffledDataset

A Shuffled a dataset.

WindowsDataset

Dataset designed to work with overlapping windows of data.

WindowsIterator

Overlapping windows iterator.

Enums

RngSource

Defines a source for a StdRng.

SizeConfig

Helper option to describe the size of a wrapper, relative to a wrapped object.

Traits

Mapper

Basic mapper trait to be used with the mapper dataset.

Window

Functionality to create a window.

Windows

Functionality to create a WindowsIterator.

Functions

iota

Generates a vector of indices from 0 to size - 1.

shuffled_indices

Generates a shuffled vector of indices up to a size.