#zarr #data #manipulating #codec #benchmark #zarrs #ome-zarr

bin+lib zarrs_tools

Tools for creating and manipulating Zarr V3 data

10 unstable releases (3 breaking)

new 0.4.2 May 16, 2024
0.4.1 May 6, 2024
0.4.0 Apr 20, 2024
0.3.0 Feb 22, 2024
0.1.1 Dec 11, 2023

#451 in Encoding

Download history 9/week @ 2024-01-22 7/week @ 2024-01-29 5/week @ 2024-02-05 109/week @ 2024-02-19 22/week @ 2024-02-26 19/week @ 2024-03-11 23/week @ 2024-04-01 179/week @ 2024-04-15 10/week @ 2024-04-22 111/week @ 2024-04-29 149/week @ 2024-05-06

449 downloads per month

MIT/Apache

260KB
6K SLoC

zarrs_tools

Latest Version msrv build

Various tools for creating and manipulating Zarr v3 data with the zarrs rust crate.

zarrs_tools is experimental and in limited production use. Correctness issues with zarrs affecting past versions of zarrs_tools are detailed here.

Changelog (CHANGELOG.md)

Tools

  • zarrs_reencode: reencode an array.
    • Can change the chunk size, shard size, codecs, fill value, chunk key encoding separator, and attributes.
  • zarrs_binary2zarr (feature binary2zarr): create an array from piped binary data.
  • zarrs_ncvar2zarr (feature ncvar2zarr): convert a netCDF variable to an array.
    • Supports multi-file datasets where a variable has been split along a single dimension.

The following tools are highly experimental and have had limited production testing:

  • zarrs_filter (feature filter): apply simple image filters (transformations) to an array.
  • zarrs_ome (feature ome): convert an array to OME-Zarr.
    • Outputs OME-Zarr 0.5-dev. This revision is currently recognised by Neuroglancer for Zarr V3.

See docs/ for tool documentation.

zarrs Benchmarking

  • zarrs_reencode: suitable for round trip benchmarking.
  • zarrs_benchmark_read_sync (feature benchmark): benchmark the zarrs sync API.
  • zarrs_benchmark_read_async (feature benchmark): benchmark the zarrs async API.

See docs/benchmarks.md for some benchmark measurements.

Install

From crates.io

cargo install --all-features zarrs_tools

From source

cargo install --all-features --path .
# cargo install --all-features --git https://github.com/LDeakin/zarrs_tools

Enabling SIMD intrinsics

Encoding and decoding performance may be improved with avx2/sse2 enabled (if supported).

This can be enabled by compiling with either of:

  • RUSTFLAGS="-C target-cpu=native"
  • RUSTFLAGS="-C target-feature=+avx2,+sse2"

Enabling non-default zarrs codecs

Non-default zarrs codecs (see zarrs crate features) can be enabled by passing them as feature flags. For example:

cargo install zarrs_tools --all-features --features zarrs/bitround,zarrs/zfp,zarrs/bz2,zarrs/pcodec

Licence

zarrs_tools is licensed under either of

Unless you explicitly state otherwise, any contribution intentionally submitted for inclusion in the work by you, as defined in the Apache-2.0 license, shall be dual licensed as above, without any additional terms or conditions.

Dependencies

~20–60MB
~1M SLoC