Dask arrays (experimental)

This module provides alternative implementations of array classes defined in the allel.model.ndarray module, using dask.array as the computational engine.

Dask uses blocked algorithms and task scheduling to break up work into smaller pieces, allowing computation over large datasets. It also uses lazy evaluation, meaning that multiple operations can be chained together into a task graph, reducing total memory requirements for intermediate results, and only the tasks required to generate the requested part of the final data set will be executed.

This module is experimental, if you find a bug please raise an issue on GitHub.

This module requires Dask >= 0.7.6.

GenotypeDaskArray

class allel.model.dask.GenotypeDaskArray(*args, **kwargs)[source]

Dask genotype array.

To instantiate from an existing array-like object, use GenotypeDaskArray.from_array().

HaplotypeDaskArray

class allel.model.dask.HaplotypeDaskArray[source]

Dask haplotype array.

To instantiate from an existing array-like object, use HaplotypeDaskArray.from_array().

AlleleCountsDaskArray

class allel.model.dask.AlleleCountsDaskArray[source]

Dask allele counts array.

To instantiate from an existing array-like object, use AlleleCountsDaskArray.from_array().