# Download from Object Storage

I have put the data onto a publically-accessible-but-not-discoverable S3 bucket.

The implications of what an S3 bucket is are not important, but it a standard way to serve objects
and is widely available.

Most OME-Zarr tools can read from S3 storages, some in a highly efficient manner.

Further, this could be used to make a tool that downloads an artefact and converts it locally.

Currently such a tool exists, except it does not support the local conversion.
The tool is developed as part of the OME (Open Microscopy Environment) project and I believe they would
be amenable to the inclusion of conversion tools in this tool.

To try it out, install uv if you do not have it already: https://docs.astral.sh/uv/getting-started/installation/

Then in a new terminal run:
```sh
uvx ome_zarr download "https://s3.g.s4.mega.io/aidjqnzhivz2k6kyxukiymjgfb4wxowhengdp/naturarv3d/CIL_basic_binned_2x2/NaturArv3D_format_viability_study/OME-Zarr/Canislupus_CZM_712__uncompressed__scales_2_4_8_16.ome.zarr"
```

This downloads the different scales of that dataset and stores them locally.

Note! This currently does not work for datasets with codecs (i.e. only works for uncompressed dataset).
This can be addressed with a small contribution to the library, which I believe they would happily take.

The workflow I imagine to get a dataset would be that the URL or full command is published on the website as part of the dataset. Then to download as a TIFF stack, the user would simply run:

```sh
uvx ome_zarr download "https://s3.g.s4.mega.io/aidjqnzhivz2k6kyxukiymjgfb4wxowhengdp/naturarv3d/CIL_basic_binned_2x2/NaturArv3D_format_viability_study/OME-Zarr/Canislupus_CZM_712__uncompressed__scales_2_4_8_16.ome.zarr" --to tiff-stack
```

The tool is maintained by the OME organisation, so we would ont have the maintenance burden associated with the tool, though we may wish to help develop this new feature.