Viktor Petukhov

I think that's the idea like [Collapse headings](https://jupyter-contrib-nbextensions.readthedocs.io/en/latest/nbextensions/collapsible_headings/readme.html) jupyter extensions.

Thanks, it works! Though to prevent Mac from blocking an unknown lib one also needs to run `xattr -r -d com.apple.quarantine telegram/lib/darwin/libtdjson.dylib`.

The latter one. I expect a constant inflow of datasets, which I want all users to see. So I'll be copying new `h5ad` objects to the `data/server` folder a lot....

Yep, I agree that refreshing on 'Load data' would be the best. Thanks for your response!

Thank you for the quick reply! Yes, I meant all variables from `adata.obs`. IMO, it's fine to leave on the user-side: if they're willing to wait, that's alright. But if...

Thank you so much!

Apologies for the late reply. Only single-channel images are supported. Would the error persist if you flatten the image to a single binary channel?

Thanks for the quick reply, Jonathan! There is no urgency from my side, and for this particular part I don't want to do sophisticated preprocessing. But otherwise, your suggestions look...

@jayp2919 , What kind of `data.object` do you use? Is it a count matrix?

Thanks for the update! In the previous version I explicitly checked that spaces are printed in the notebook. Indeed, huge notebook size was exactly the reason I started looking on...