xzarrguard 0.1.1

Integrity checks and creation helpers for Zarr v3 stores

xzarrguard

xzarrguard provides concise APIs and a CLI to validate completeness of local Zarr v3 stores and create stores with explicit no-data policy.

Install

PyPI: pip install xzarrguard
conda: conda install xzarrguard
from source: pip install .

Install-free CLI usage

uv: uvx xzarrguard check /path/to/store.zarr
pixi: pixi exec xzarrguard check /path/to/store.zarr

API quickstart

from xzarrguard import check_store, create_store

report = check_store("store.zarr")
if report:
    print("store is complete")

create_store(
    dataset,
    "store.zarr",
    no_data_chunks={"temperature": [(0, 0)]},
    no_data_strategy="manifest",
)

Write and guard in one step (wrapper around .to_zarr()):

from xzarrguard import guarded_to_zarr

guarded_to_zarr(dataset, "store.zarr")

In-place metadata-only guard update (no chunk rewrite):

create_store(
    None,
    "store.zarr",
    no_data_chunks={"temperature": [(0, 0)]},
    in_place_metadata_only=True,
)

Treat the current store as baseline and derive allowed-missing chunks from what is currently missing:

create_store(
    None,
    "store.zarr",
    in_place_metadata_only=True,
    infer_no_data_from_store=True,
)

CLI quickstart

xzarrguard check store.zarr
xzarrguard create source.zarr target.zarr --no-data no_data.json
xzarrguard create store.zarr --in-place-metadata-only --no-data no_data.json
xzarrguard create store.zarr --in-place-metadata-only --infer-no-data-from-store

Coverage

pytest

pytest prints terminal coverage and writes coverage.xml.

Documentation

https://j-haacker.github.io/xzarrguard/

zensical serve
zensical build --clean

Channels

• PyPI: https://pypi.org/project/xzarrguard/
• Conda-forge: https://anaconda.org/conda-forge/xzarrguard
• GitHub: https://github.com/j-haacker/xzarrguard

Release (maintainers)

# bump src/xzarrguard/_version.py first
python -m build
python -m twine check dist/*
python -m twine upload dist/*

Use a PyPI API token for upload (for example TWINE_USERNAME=__token__). For conda-forge, update recipe/recipe.yaml after the PyPI release (fixed version + PyPI sdist URL + sha256), then submit a recipe/feedstock PR.

Acknowledgement: Initial scaffolding and implementation assistance by OpenAI Codex.