zarr-python

Zarr is a Python implementation of the Zarr storage format, providing compressed, chunked, N-dimensional arrays that work well for parallel computing and object storage. It supports creating arrays with NumPy-compatible dtypes, chunking along any dimension, and encoding chunks with various compressors. Arrays can be stored in memory, on disk, inside zip files, or on object storage such as S3. The library supports concurrent reading and writing from multiple threads or processes, and arrays can be organized into hierarchical groups. Installation is available via pip or conda. The repository also includes sub-packages for metadata tools, lazy indexing of chunked arrays, and an HTTP server targeting Zarr data.