I think this should be added to the validation scripts, to check that the atlases have additional_references set correctly (or at all).

The warning is useful in this case, I think, @viktorpm

UserWarning: This atlas seems to be outdated as no additional_references list is found in metadata!

This suggests that there is an atlas (is it the human atlas or the one after in the validation list?) that hasn't been updated since we introduced the additional references concept in the code.

How I think it is intended to work (for atlases with no additional references):

• the atlas packaging script calls the wrapup function without specifying the additional_references keyword argument e.g. hte human one.
• the wrapup function will therefore use the default additional_references = {} and will have an empty dictionary.

@alessandrofelder, looks like the human atlas doesn't return what the validate_additional_references function expects. The warning has always been there, so that shouldn't be a problem. I tried to print the reference list allenhu.additional_references.references_list but got an error. Should we try to catch the error and continue the validation or fix the atlas (it is likely a bigger job)?

allen mouse 100um atlas

allen100 = BrainGlobeAtlas("allen_mouse_100um") 

allen100.additional_references.references_list

[]

allen human 500um atlas

allenhu = BrainGlobeAtlas("allen_human_500um")
/home/plattnerv/miniforge3/envs/bg-atlasapi/lib/python3.11/site-packages/brainglobe_atlasapi/core.py:65: UserWarning: This atlas seems to be outdated as no additional_references list is found in metadata!
  warnings.warn(

allenhu.additional_references.references_list

Traceback (most recent call last):
  File "/home/plattnerv/.pycharm_helpers/pydev/pydevconsole.py", line 364, in runcode
    coro = func()
           ^^^^^^
  File "<input>", line 1, in <module>
AttributeError: 'BrainGlobeAtlas' object has no attribute 'additional_references'

To fix the issue, I think (famous last words) that the human atlas just needs to be regenerated. It should then have the correct metadata. I think the issue stems from the fact that it hasn't been regenerated since the concept of additional references was added to the atlas API.

I agree, but I couldn't find an easy fix for the script. I'll make it my priority to work on that (in a new PR).