Please find here some minimal information to start with s3 storage and usage.
S3 storage is optimzed for small chunck. So if you have gigantic time-lapse, z-stack, multi-channels it is recommended to convert it, store it as a ome-ngff file .
This will take time as all the files permissions as to be set.
conda create ome_ngff
conda install napari[all]
Here is a minimal example using a dataset from IDR
import dask.array as da
import napari
path = "https://uk1s3.embassy.ebi.ac.uk/idr/zarr/v0.4/idr0101A/13457227.zarr/0"
dask_data = da.from_zarr(path)
viewer = napari.Viewer()
new_layer = viewer.add_image(dask_data )
To work with a dataset on s3.epfl.ch , I had to modifiy permissions (read) granted to All Users and Any AWS Users.
Note that :
So "strangers" should not access your data 🎉
import dask.array as da
import napari
path = "https://s3.epfl.ch/12092-127a2c5216de10513f28977152b492d9/Test_zarr/ome-tiff.companion.ome_lvl_0.ome.zarr/0/"
dask_data = da.from_zarr(path)
viewer = napari.Viewer()
new_layer = viewer.add_image(dask_data )