Class methods' full source path is shown instead of the identifier used for it

[ianspektor]

Describe the bug

Hey! Thank you very much for the work on #90 and #91 :)

91 does seem fixed, but 90 doesn't. Upgraded mkdocstrings-python to 1.5.3 and griffe to 0.36.8 (see PR) and I'm still seeing the object's full path (temporian.implementation.numpy.data.event_set.EventSet.begin) instead of the identifier I used for it (::: temporian.EventSet.begin) in the header (see the rendered doc here). Note that I import the EventSet class in my root __init__ which makes the identifier work.

Full explanation in the original issue.

To Reproduce
See #90

Expected behavior
See #90

Thanks in advance for any help!

[pawamoy]

Hmm, weird, I expose objects from submodules in my root init module too, and they correctly keep their alias path in the docs, for example: https://mkdocstrings.github.io/griffe/reference/griffe/#griffe.Attribute.

I'll check out your repo again 🙂

[pawamoy]

Actually I'm able to replicate. It works fine when I let mkdocstrings unfold members of griffe or griffe.Attribute, but if I inject griffe.Attribute.functions directly, the path is the canonical one.

[ianspektor]

Great to hear! Let me know if there's anything you need from my side

[pawamoy]

Pushed a fix, will release soon 🙂