Skip to content

Installing yaozarrs#

The basic package (with no extras) supports metadata creation & validation, but lacks structural validation and writing capabilities.

pip install yaozarrs

uv add yaozarrs

To install the bleeding-edge development version from GitHub, (shown here using the io extra as an example):

pip install "yaozarrs[io] @ git+https://github.com/imaging-formats/yaozarrs.git"

conda install conda-forge::yaozarrs

Note that the conda distribution contains only the core package and no extras. You will need to manually install the following dependencies if you need them:

  • conda-forge::fsspec - for yaozarrs.open_group and any structural validation.
  • conda-forge::zarr OR conda-forge::tensorstore - for writing support.

Structural validation#

To enable validation of Zarr hierarchies, include the io extra when installing. This brings in fsspec (but not zarr-python):

pip install "yaozarrs[io]"

uv add "yaozarrs[io]"

Writing support#

If you want to use the convenience functions in yaozarrs.write, to create complete OME-Zarr stores with array data, you will need to pick a zarr-array library:

  • zarr-python#

    zarr-python is the reference implementation of the Zarr specification for Python.

    pip install "yaozarrs[write-zarr]"

    uv add "yaozarrs[write-zarr]"

    Then use writer="zarr" in the yaozarrs.write functions.

  • tensorstore#

    Tensorstore is an alternative zarr-array library developed by Google, which can offer better I/O performance in most cases.

    pip install "yaozarrs[write-tensorstore]"

    uv add "yaozarrs[write-tensorstore]"

    Then use writer="tensorstore" in the yaozarrs.write functions.

  • Custom backend#

    If you don't want to use either of the provided backends, you can implement your own array-writing functionality