Feature request: fusion as zarr with sharding (zarr3) #81
Description
This would be a welcome feature as it would simplify the subsequent handling of multi-TB fused datasets: no more humongous number of files. I understand that that the zarr3 specification is stable now and is being implemented in several libraries and tools. I have no idea how BigStitcher accesses saving as zarr, but perhaps it could integrate directly from one of those updated libraries?