on disk reading of a large h5ad file

[hoondy]

I am working with a large h5ad file and it requires very large memory to load the data using pg.read_input. AnnData allows users to read an h5ad file partially on memory (.obs or .var) and leave the count data (.X) on disk.

See https://anndata-tutorials.readthedocs.io/en/latest/getting-started.html#Partial-reading-of-large-data
See also https://anndata.readthedocs.io/en/latest/fileformat-prose.html

I was wondering if something similar can be done using pegasus. Single-cell data is getting larger and I feel like this should be a useful option for users working with a limited amount of memory.