argopy.stores.ArgoKerchunker.translate

ArgoKerchunker.translate(ncfiles: str | Path | List, fs=None, chunker: Literal['first', 'auto', 'cdf3', 'hdf5'] = 'first')

Translate netcdf file(s) into kerchunk reference data

Kerchunk data are saved on the ArgoKerchunker instance store.

Once translated, netcdf file reference data are internally registered in the ArgoKerchunker.kerchunk_references attribute.

If more than 1 netcdf file is provided, the translation is executed in parallel:

• if Dask is available we use dask.delayed/dask.compute(),

• otherwise we use a concurrent.futures.ThreadPoolExecutor.

Parameters:

• ncfiles (Union[str, Path, List]) – One or more netcdf files to translate

• fs (None) – An argopy file store, inheriting from ArgoStoreProto.

• chunker (Literal['first', 'auto', 'cdf3', 'hdf5'] = 'first') –

  Define the kerchunker formater to use. Two formater are available: kerchunk.netCDF3.NetCDF3ToZarr or kerchunk.hdf.SingleHdf5ToZarr:

  • ’first’: detect and select formater from the first netcdf file type

  • ’auto’: detect and select formater for each netcdf of the nc files

  • ’cdf3’: impose use of kerchunk.netCDF3.NetCDF3ToZarr for all nc files

  • ’hdf5’: impose use of kerchunk.hdf.SingleHdf5ToZarr for all nc files

Return type:

List(dict)

See also

ArgoKerchunker.to_reference()