Refactor built-in plans to use contingency_wrapper

[danielballan]

The contingency_wrapper is better than the finalize_wrapper because the "final plan" takes in the exception that was raised as the argument, which allows the clean up to target specific failure modes if desired.

Even where that capability isn't used, it would be an improvement I think to model using contingency_wrapper in our code so that users can easily discover it and find examples of its use.

[callumforrester]

Reviewers: @danielballan @tacaswell

[stan-dot]

should mark finalize_plan as deprecated?

docs/tutorial.rst tells the user to use finalize_wrapper, should that be changed?

@danielballan

[stan-dot]

files that mention finalize_wrapper

• preprocessors.py
• /tests/test_generators.py - testing specifically the function itself
• /tests/test_new_examples.py - also in test_finalize
• /tests/test_run_engine.py - the most clear location where it might be replaced

When I'm looking atthe plan_patterns.py I do not see any wrappers.
OTOH, plan_stubs.py contains references to wrappers, and those used there are: rewindable, contingency and relative_set.

I am not sure where the change is to be made.

[tacaswell]

I'm weakly against deprecating it. In cases where you only want to do cleanup the simpler signature is nice.

I'm also sure we have some usage on the floor at facilities who have already deployed bluesky and the disruption of deprecating this does not seem worth the marginal gain of getting rid of name in the namespace.

On the other hand "pause for debug" was a bad idea (I think I wrote it).

Maybe we should only purge it from the narrative docs?

[prjemian]

Not seeing any current use of finalize_wrapper, finalize_plan, or contingency_wrapper in our standard kit. An APS beam line might be using any of these, but not likely.

[stan-dot]

should the issue be closed as not needed then?
@danielballan @prjemian

[prjemian]

The action for this issue might be to ensure the documentation for the various affected code contains the recommendation in the initial statement here:

The contingency_wrapper is better than the finalize_wrapper ...

[prjemian]

@stan-dot -- Thanks for picking this one up!

[stan-dot]

no worries! doing a PR for the recommedation for the edit now

[stan-dot]

so far only the tutorial.rst got modified with the menion, I'm happy to merge as is unless more places need the change