Skip to main content

Crate openreadout_zarr

Crate openreadout_zarr 

Source
Expand description

OME-Zarr (OME-NGFF) reader: NGFF 0.1–0.5 on Zarr v2 and v3, directory and zip stores.

Images come from multiscales groups, HCS plates (one image per field of view) and bioformats2raw.layout collections (one image per series, with OME/METADATA.ome.xml parsed by the TIFF crate’s OME-XML parser). Arrays are read through the pure-Rust zarrs crate; the blosc, zstd and numcodecs lz4 codecs are decoded by openreadout-codecs (see [codecs]) so that no C library is needed.

Layout and vocabulary: docs/formats/ome-zarr.md. Provenance: docs/provenance/ome-zarr.md.

§Example

Every reader implements openreadout_core::FormatReader: descriptor says what it reads, sniff recognises a file from its first bytes, and open returns a openreadout_core::Dataset with a header-only info() summary and lazy data access.

use openreadout_core::FormatReader;
use openreadout_zarr::ZarrReader;

let format = ZarrReader.descriptor();
assert_eq!(format.id, openreadout_zarr::FORMAT_ID);
println!("{} ({}): .{}", format.name, format.family, format.extensions.join(", ."));

Reading a file:

use std::path::Path;

use openreadout_core::{FormatReader, PlaneIndex};
use openreadout_zarr::ZarrReader;

let mut dataset = ZarrReader.open(Path::new("plate.ome.zarr"))?;
let info = dataset.info()?; // headers only: fast even on multi-gigabyte files
for image in &info.images {
    println!(
        "image {}: {} x {} px, {} channel(s), {} plane(s)",
        image.index, image.size_x, image.size_y, image.size_c, image.plane_count
    );
}
// Pixels are read one plane at a time: image 0, channel 0, first Z slice, first time point.
let plane = dataset.read_plane(0, PlaneIndex { c: 0, z: 0, t: 0 })?;
assert_eq!(plane.data.len(), plane.expected_len());

Applications that accept any instrument file usually register every reader in an openreadout_core::Registry and let it detect the format; that is what the openreadout command-line tool does.

§API stability

The supported API is what this page documents. The parser modules are public only so that tests and fuzz targets can reach them: they are hidden from this documentation and may change in any release.

Structs§

ZarrReader
The OME-Zarr reader.

Constants§

FORMAT_ID
Format id used on the command line and in JSON.