list atlases

[FedeClaudi]

This PR introduces functionality to list all available and downloaded brainglobe atlases, these get printed to console as such:

this can be done by either calling brainatlas_list_atlases in the terminal or with:

import brainglobe_api
brainglobe_api.list_atlases()

limitation: currently the list of 'available' atlases is specified manually and it uses the classes defined in brainglobe_api.bg_atlas.py, it would be nicer to have a more automated way of doing this.

[vigji]

maybe better to use glob and a pattern to match

[FedeClaudi]

I haven't used pathlib much so not very familiar with best practices...will fix

[FedeClaudi]

With the latest changes I've fixed a few tests (hopefully) and added brainglobe_api.get_atlas_class_from_name which does what the name says.

Unless there's anything left to change/fix I think we can merge.

[adamltyson]

This PR introduces functionality to list all available and downloaded brainglobe atlases, these get printed to console as such:

Awesome, I love this

[vigji]

There's still 3 things that might probably be good to solve here:

1. fix the atlases classes naming inconsistency: I think all should have or not species name, all should have or not Brain, and all should have or not Atlas. I don't strongly mind any of those choices, as long as they are all consistent. I would go for having species, not having brain, and having atlas;
2. There's no point in keeping the version in the class anymore, that we should probably drop and find automatically the version from the remote atlas name - although I am not sure how to fetch it, but there should be a way. Keeping these specifications in multiple places is redundant and error-prone. But I can do it myself in a different PR if you want!
3. I think it would be good to show the original source reference for each atlas

[vigji]

Probably this could be handled with something like what is decribed here:
https://stackoverflow.com/questions/1796180/how-can-i-get-a-list-of-all-classes-within-current-module-in-python

Putting some conditions on the class name (eg it should not be BGAtlas)

[FedeClaudi]

Sure we can do that as well, I've done it manually because we don't have that many and they won't be changing too much.

[FedeClaudi]

regard naming consistency, I agree it would be good to do that.
My considerations:

• we can skip Brain from the name as they are all brain atlases
• we can remove the version, and maybe have it as a class attribute?

I propose for the name: [name]_[species]_[atlas]_[resolution] with name distinguishing for example between the Allen and the Kim mouse atlas.

I think it would be good to show the original source reference for each atlas

When printing the atlases list? The problem with that is that I couldn't create an instance of e.g. FishAtlas without downloading the stuff so not sure how to access the reference metadata without downloading the atlases? If there's a way to just get the metadata then sure.

Regardless we should have a README somewhere (github, gin, both) with a table of all atlases and some basic info (species, resolution, source...)

[adamltyson]

we can skip Brain from the name as they are all brain atlases

I would rather keep the brain in the name. Theoretically this could be used with data from other organs.

[FedeClaudi]

@vigji do you think we can merge this PR and deal with class names separately? This way I can finalise the 3d viewer stuff :)

[vigji]

When printing the atlases list? The problem with that is that I couldn't create an instance of e.g. FishAtlas without downloading the stuff

Maybe we could show it at least for the downloaded stuff while we figure tout the point below.

Regardless we should have a README somewhere (github, gin, both) with a table of all atlases and some basic info (species, resolution, source...)

I agree that would be convenient to have! Maybe in the default folder.

But yes, as this merge is required, let's merge this and we can polish these things in a separate PR!