chore: move to zarr

[ilan-gold]

Turns out most of tutorials are using zarr anyway!

[review-notebook-app]

Check out this pull request on 

See visual diffs & provide feedback on Jupyter Notebooks.

---

Powered by ReviewNB

[Zethson]

One row per sentence please to make this more reviewable.

[Zethson]

1. A getting started tutorial should not have warnings like

/var/folders/k9/9wc7lvwj2g34_r74kn6cr0nr0000gn/T/ipykernel_12269/2319192822.py:5: FutureWarning: `__version__` is deprecated, use `importlib.metadata.version('anndata')` instead.
  print(ad.__version__)

and

/var/folders/k9/9wc7lvwj2g34_r74kn6cr0nr0000gn/T/ipykernel_12269/2822107891.py:2: FutureWarning: Automatic reshaping when setting X will be removed in the future.
  adata[:3, 'Gene_1'].X = [0, 0, 0]
/var/folders/k9/9wc7lvwj2g34_r74kn6cr0nr0000gn/T/ipykernel_12269/2822107891.py:2: ImplicitModificationWarning: Setting element `.X` of view, initializing view as actual.
  adata[:3, 'Gene_1'].X = [0, 0, 0]

2. Maybe we should have a dropdown that explains more details about hdf5 & zarr and when to use what?

[ilan-gold]

Maybe we should have a dropdown that explains more details about hdf5 & zarr and when to use what?

I added a bit more, but not sure why it should be a dropdown since the info is important. There's so much to say but I tried to wrap it up a bit into the main points - parallelism and async.

[Zethson]

Nit: This is still on a single line:

Read/Write operations on h5ad and Zarr are supported. One should note that the lazy objects are materialized when this is called. For now, the anndata loaded from file won't be loaded with dask arrays in it. zarr is generally much faster due to the lack of a read-lock.