Skip to content

Latest commit

 

History

History
 
 

zarr

README.md

zarr data source

The "zarr" data source allows Neuroglancer to directly read zarr format arrays, using the following data source URL syntax:

zarr://FILE_URL, where FILE_URL is a URL to the directory containing the .zarray (v2) or zarr.json (v3) metadata file using any supported file protocol.

Alternatively, FILE_URL may be a URL to the directory containing the .zattrs (v2) or zarr.json (v3) group metadata file that specifies an OME-NGFF multiscale dataset. Multiscale metadata versions 0.4, 0.5-dev and 0.5 are supported.

Supported data types (little and big endian):

  • uint8
  • int8
  • uint16
  • int16
  • uint32
  • int32
  • uint64
  • float32

As an extension, dimension units may be specified using the dimension_units user attribute. For example, to specify that the voxel size is 4x5x30nm for a 3-d array, the following attribute may be set:

{
  "dimension_units": ["4 nm", "5 nm", "30 nm"]
}

Zarr v2

If the zarr array uses / rather than the default of . as the dimension separator in chunk keys, you can either specify the separator as the dimension_separator member in the .zarray metadata file (preferred) or use a data source URL of zarr://FILE_URL?dimension_separator=/.

Supported compressors:

  • blosc
  • gzip
  • null (raw)
  • zlib
  • zstd

Filters are not supported.

Dimension names may be specified using an _ARRAY_DIMENSIONS attribute, as defined by xarray.

Zarr v3

Supported codecs:

  • crc32c (not validated)
  • blosc
  • bytes
  • gzip
  • sharding_indexed
  • transpose
  • zstd