uncompounded profiles break viz
jamie256 opened this issue · comments
Jamie Broomall commented
Description
we uncompound metrics so that they map to feature charts in WhyLabs, but if you download that profile again, its hard to compare these profiles to local profiles in the case of images.
Also if you view the to_pandas() of an uncompounded profile, its very confusing because the table would then look sparse.
Suggestions
- Be able to detect uncompounded state and do that as needed in notebook profile viz
- make to_pandas() show something reasonable on uncompounded profiles
- I have reviewed the Guidelines for Contributing and the Code of Conduct.
github-actions commented
This issue is stale. Remove stale label or it will be closed next week.