I would like to load a HEST object without pooling and without creating spots for Xenium data. Essentially, I'd like to work directly with the raw transcript data, with each cell as an individual observation.

From issue #79, I understand that this is possible by accessing the raw transcript data directly. However, doing so means I can't use the standard Scanpy functions (based on cells).

Is there currently a supported way to obtain a HEST object with AnnData object containing n_obs == num_cells instead of num_spots?