[uss_qualifier/astm] Make cleanup phase in down_uss use estimated extents as well

[the-glu]

This PR should fix #1400 by also using estimated extends during cleanup phase.

Initial cleanup was not doing anything during initial phase since it was using scenario_execution_max_extents only, where as check was done on estimated_max_extents.

I didn't verify exact scenario in original issue, but looking at report, there is now a query to perform clean up:

(before, no query was done:

so potentially existing OP intents would not being deleted

)

And having a previous test leaving an OPI cleanup it as expected:

The other usage of estimate_scenario_execution_max_extents don't perform any cleanup (but does perform an 'area clear' check), so I didn't change anything there.

[BenjaminPelletier]

This is a pre-existing problem, but we're initiating the check too early. Ideally, we should complete the action to collect data before we initiate the check to evaluate that data. But, it would be inconvenient to need to repeat the with initiation statement in two different places, so I think practically it's ok to initiate the check before performing the action (find_op_intent) since starting to perform the action necessarily means we're going to perform the check one way or another. But, we should not initiate the check if we're not actually going to perform the action (and therefore definitely not perform the check).

If area is None, currently the check will indicate as passing, but that is incorrect as we haven't checked anything. Instead, we should make sure we're actually going to check something before initiating the check since initiating a check and not recording a failure is defined as meaning that the check has positively passed.

[the-glu]

Yes, since it's now back to a single, valid area it should be better (at least not worse)