Rarr: A native R reader for Zarr
Author(s): Mike L Smith
Affiliation(s): European Molecular Biology Laboratory
Social media: https://twitter.com/grimbough
The Zarr file format has been developed for the storage of large multi-dimensional arrays. Its design is specifically focused to facilitate easy access to datasets stored in the cloud, and this has led to adoption of the Zarr file format across a wide range of scientific disciplines, including genomics, astrophysics, earth sciences and microscopy imaging. Despite this rapid growth in popularity for the Zarr format, adoption by Bioconductor packages has been hindered by a lack of support for the format in R, with limited accessibility provided by packages that wrap tools provided in other languages. Here I introduce Rarr, an R package that supports the Zarr specification without reliance on third-party tools. Rarr supports the majority of datatypes defined in the Zarr specification, as well as a variety of compression tools and data layouts, allowing interoperability with other languages that support the format. Additionally, Rarr is able to access Zarr arrays both on local storage and in the cloud.