Closed danielballan closed 4 years ago
Rebased on #1 (which I just merged). We should add HDF5 handler tests as part of this PR. There are currently none at all.
Good point, can't think of a reason not to do that. Fix pushed, will merge if CI passes.
We previously relied on PIMS do this. To avoid some of the assumptions baked into PIMS, I think that (at least for now) we need format-specific solutions. I am starting with h5py and ignoring TIFF because TIFFs tend not to be too large for memory, so it is not as critical to make them lazy.
The other h5py handler(s) should be updated similarly before this is merged.