Bigtable admin: helper to await an optimized restore table #7210

Objective

Help our Bigtable customers determine when a restored table is optimized and ready for use.

Background

We have a GAPIC LRO for bigtable table admin's RestoreTable RPC. This operation completes when a table is usable, but not necessarily when the table is best used.

There is some optimize_table_operation_name field in the response that tracks a secondary LRO for when the table is optimized. Applications that want to wait until the restored table has been optimized have to manually poll this operation name with GetOperation and extract the proper type in the response Any.

We should expose some sort of first-class way for applications to await the secondary optimize LRO.

Design Sketch

I would probably add some new type like RestoredTablePoller or something and extend the builder::RestoreTable to return it, similarly to what we do with BQ's into_job_poller().

I would consider just adding an until_optimized() on the top level builder... but (nit) I think this does not let an application take a restored table and save the operation name across processes. 🤷

There is also a nuance around: we are performing two LRO loops. Do we need a separate policy for each one?