Some numerical attributes are shown as text and vice versa

[slinnarsson]

In Cell Metadata view, some attributes that contain only floats ("ClassProbability...") are shown as text, and some that contain only text ("Class") are shown as numbers(*):

(*) Actually, Class is an attribute that contains only strings, and they are all "Neurons". This is for some reason shown as a single 1.

[JobLeonard]

I don't think the first one is a bug: For whateve reason, Class has only one unique value shared with every attribute. So the viewer assumes it is more understandable to only list that value.

The other attributes have unique values for each value, and we cannot list all of those. So instead only the first five are shown. I think the original reasoning was to apply this to string attributes with unique markers (for example, Genes or CellID - although ironically those are often not shown like this in my testing datasets because there usually are a handful of duplicates strings). However, these are numerical values, so that should change.

[slinnarsson]

But why is the single value Neurons shown as 1? I would expect it to be shown as Neurons.

[JobLeonard]

Ah, I see what you mean (did not know the value should be Neurons instead of 1). Looks like a simple bug: I forgot to check for indexed strings there. Will fix, give me a minute

[JobLeonard]

Should be fixed:

[JobLeonard]

Should I switch over to Box as the default over Bar?

[JobLeonard]

[slinnarsson]

yes, sure.