Inferencedata warning

[Pyrsos]

Update sample method inferencedata warning, to include specific type and also define the stacklevel parameter of the warning message. This addresses issue #4481

Depending on what your PR does, here are a few things you might want to address in the description:

• what are the (breaking) changes that this PR makes?
• important background, or details about the implementation
• are the changes—especially new features—covered by tests and docstrings?
• linting/style checks have been run
• consider adding/updating relevant example notebooks
• right before it's ready to merge, mention the PR in the RELEASE-NOTES.md

[Spaak]

Hi @Pyrsos, thanks for the PR! At first sight, the failing test (on Windows only?) appears unrelated to the change here. Any thoughts? I'll have a look as well.

[Spaak]

Could you rebase your PR onto current master? There are some conflicts now; once you rebase (or merge current master into your branch) I should be able to merge this, and hopefully the failing test script will disappear as well.

[Spaak]

There were a few commits related to bash scripts etc. (undone later) that had nothing to do with this PR; I've now rebased onto current master and picked only the relevant commits. Let's see what CI shows.

Rebased

[Pyrsos]

@Spaak Sorry, I think I messed this up. Tried to rebase and I think I reintroduced all the other commits.

[Spaak]

@Spaak Sorry, I think I messed this up. Tried to rebase and I think I reintroduced all the other commits.

Yep :D No problem, definitely fixable. I already did the rebasing, might be good if you do it instead. What I did was:

gh pr checkout 4493 (which you wouldn't need to do as you have the branch locally already I assume)
git rebase -i master

In the interactive rebase, be sure to only pick those three commits actually related to this change. Drop the rest.

Then, git push -f for a force-push, which will overwrite the history here in the PR branch to the new, clean, one.

[Spaak]

Hmm the old commits are still there. Is your master up to date? (git checkout master and git pull upstream master before doing the rebasing)

[Pyrsos]

Hmm the old commits are still there. Is your master up to date? (git checkout master and git pull upstream master before doing the rebasing)

Yeah, I have the latest master. Not sure what is going on exactly...

[Spaak]

You may have had a merge commit locally on master (created after pulling upstream could not be done in a fast-forward), and therefore rebased onto that merge commit. I've now rebased again (excluding the merge commit), leaving only the three key commits.

If you want to get this branch locally in a clean state (though you might not need it if the tests succeed and we merge into master), it's easiest to delete it locally and fetch again.

[Pyrsos]

@Spaak OK, I think I figured it out. I had messed up the master branch on my local repo. Sorry again for all the hustle. It looks like it has the correct commits now.

[Spaak]

Haha and both of us now did the same thing :) In any case it looks clean now.

[Pyrsos]

Haha and both of us now did the same thing :) In any case it looks clean now.

Yeah, looks that way! Thanks for the help! Hopefully it builds now!

[MarcoGorelli]

I usually do

git fetch upstream
git rebase -i upstream/master

Because even if I messed up my local master branch, upstream/master is probably fine :)

[michaelosthege]

The test failure is the flaky test (#4323).
https://github.com/pymc-devs/pymc3/pull/4493/checks?check_run_id=2004362931#step:7:2461

Just re-start the Github action pipeline and it will probably go away.

[twiecki]

Thanks @Pyrsos!