What are the zarr heuristics used for chunking in to_zarr()?

[hansmohrmann]

From the documentation in xarray.Dataset.to_zarr(), if no chunks are provided,

If neither Dask chunks nor encoding chunks are present, chunks will be determined automatically by Zarr

If not other chunks are found, Zarr uses its own heuristics to choose automatic chunk sizes

Is there a way to get these chunks before calling to_zarr() (without explicitly assigning chunks)? I'm assuming that this is done by zarr and not xarray.

Thank you very much!

[d-v-b]

i believe that logic is inside this function:

zarr-python/zarr/util.py

Line 81 in e1b71a3

def guess_chunks(shape: Tuple[int, ...], typesize: int) -> Tuple[int, ...]:

[hansmohrmann]

Thank you, that's perfect! zarr.util.guess_chunks(da.shape, typesize=da.dtype.itemsize) got me there.